From 7fa124de5e6eccf1f7ee84d911013be368c2eabc Mon Sep 17 00:00:00 2001 From: Neil <4138956+nwparker@users.noreply.github.com> Date: Tue, 10 Nov 2026 16:19:45 -0800 Subject: [PATCH 01/21] relay edge: silent host at authorization expiry closes 4408; dead-control diagnostics (#40929) * relay edge: a keepalive host already silent when its authorization expires closes 4408 At token exp + 60 s, a keepalive control that went silent (asleep, offline) had its silence probe pending, so the expiry check closed it 4401. Production showed these as 4401 'never refreshed' closes with the host silent for ~5-6 minutes, inflating the auth-failure count and hiding drops from the 4408 comparison with the cells. Such a host now closes 4408 (orphan time back-dated as for any silence close). Local e2e K4 (fails without the change: 4401). 48/48. * relay edge: handshake time on control_active, control age and silence on dead-control closes At 1%, many new controls went silent at or right after activation and were closed minutes later, with no rebound from a retry. To tell a desktop that gave up the connect (15 s deadline, e.g. a slow first hello on a new object) from one that died later: control_active double2 = ms from the upgrade to activation; new silence_close (seconds silent, where found, keepalive/legacy, age); auth_expired_close double3 = control age. --- cloud/apps/relay-edge/src/host-relay-base.ts | 3 +++ .../apps/relay-edge/src/host-relay-control.ts | 10 ++++++- .../relay-edge/src/host-relay-handshake.ts | 3 ++- cloud/apps/relay-edge/src/host-relay.ts | 15 +++++++++-- cloud/apps/relay-edge/src/metrics.ts | 8 ++++-- .../relay-edge/test/harness-mvp-object.ts | 4 +-- .../relay-edge/test/mvp-object-hardening.ts | 26 +++++++++++++++++++ cloud/apps/relay-edge/test/mvp-object.ts | 4 +-- 8 files changed, 63 insertions(+), 10 deletions(-) diff --git a/cloud/apps/relay-edge/src/host-relay-base.ts b/cloud/apps/relay-edge/src/host-relay-base.ts index f3f34839b10..d550b6ca840 100644 --- a/cloud/apps/relay-edge/src/host-relay-base.ts +++ b/cloud/apps/relay-edge/src/host-relay-base.ts @@ -45,6 +45,9 @@ export type ControlAttachment = { clientKeepalive?: boolean // A keepalive control found silent: one relay ping sent at this time (see probeSilentControls). probeAt?: number + // Upgrade and activation times, for metrics (absent on attachments from before these fields). + openedAt?: number + activeAt?: number } export type Attachment = ControlAttachment export type Conn = { diff --git a/cloud/apps/relay-edge/src/host-relay-control.ts b/cloud/apps/relay-edge/src/host-relay-control.ts index 16697097b1c..2cc2756c22c 100644 --- a/cloud/apps/relay-edge/src/host-relay-control.ts +++ b/cloud/apps/relay-edge/src/host-relay-control.ts @@ -28,8 +28,16 @@ export abstract class HostRelayControl extends HostRelayHandshake { // Where the host is going away anyway (export, erase, drain, a new control of the same host): a // silent control is closed now. Its orphan time is the latest the ping alarm would have found it, // so rebind and the orphan grace are never shorter than before. + // A control closed for silence: value = seconds silent, label = where it was found, reason = + // keepalive | legacy; double2 = seconds since it became active (-1 if unknown). + protected deadControlMetric(a: ControlAttachment, now: number, seen: number, found: 'alarm' | 'on_demand' | 'expired'): void { + const ageS = a.activeAt === undefined ? -1 : Math.round((now - a.activeAt) / 1000) + metric(this.env, 'silence_close', Math.round((now - seen) / 1000), found, [ageS], this.keepalive(a) ? 'keepalive' : 'legacy', a.place) + } + protected async closeSilentControls(now = Date.now()): Promise { - for (const { ws, seen } of this.silentControls(now)) { + for (const { ws, a, seen } of this.silentControls(now)) { + this.deadControlMetric(a, now, seen, 'on_demand') this.close(ws, RELAY_CLOSE_CODE.PEER_DROPPED, 'control silence timeout') await this.onSocketGone(ws, RELAY_CLOSE_CODE.PEER_DROPPED, '', Math.min(now, seen + this.config.silenceTimeoutMs + this.config.pingIntervalMs)) } diff --git a/cloud/apps/relay-edge/src/host-relay-handshake.ts b/cloud/apps/relay-edge/src/host-relay-handshake.ts index 22614a1fe85..1d9995d44bf 100644 --- a/cloud/apps/relay-edge/src/host-relay-handshake.ts +++ b/cloud/apps/relay-edge/src/host-relay-handshake.ts @@ -157,6 +157,7 @@ export abstract class HostRelayHandshake extends HostRelayBase { }) const leaseExpiresAt = now + LEASE_BASE_MS + Math.round((Math.random() * 2 - 1) * LEASE_JITTER_MS) a.phase = 'active' + a.activeAt = now a.generation = generation a.lastPongAt = now a.lastPingAt = now @@ -191,7 +192,7 @@ export abstract class HostRelayHandshake extends HostRelayBase { }) // After the ack, which strict shipped parsers must see unchanged; only hosts that asked get it. if (a.clientKeepalive) { this.send(ws, { type: 'keepalive-offer', v: 1, intervalMs: this.config.pingIntervalMs }) } - metric(this.env, 'control_active', 1, rebind ? 'rebind' : 'new', [], a.clientKeepalive ? 'keepalive' : 'legacy', a.place) + metric(this.env, 'control_active', 1, rebind ? 'rebind' : 'new', [a.openedAt === undefined ? -1 : now - a.openedAt], a.clientKeepalive ? 'keepalive' : 'legacy', a.place) await this.reschedule() } } diff --git a/cloud/apps/relay-edge/src/host-relay.ts b/cloud/apps/relay-edge/src/host-relay.ts index 14f5a0f544f..bc6095bc33a 100644 --- a/cloud/apps/relay-edge/src/host-relay.ts +++ b/cloud/apps/relay-edge/src/host-relay.ts @@ -145,7 +145,8 @@ export class HostRelay extends HostRelayPhones { claims, origin, caps: [...parseRelayHostCapabilities(request.headers.get('x-orca-host-capabilities') ?? '')], - deadline: Date.now() + RELAY_PROTOCOL_LIMITS.firstFrameDeadlineMs + deadline: Date.now() + RELAY_PROTOCOL_LIMITS.firstFrameDeadlineMs, + openedAt: Date.now() }) await this.reschedule() return new Response(null, { status: 101, webSocket: client }) @@ -180,15 +181,25 @@ export class HostRelay extends HostRelayPhones { continue } if (now - seen > this.config.silenceTimeoutMs && (!this.keepalive(a) || now - a.probeAt! >= KEEPALIVE_PROBE_MS - 50)) { + this.deadControlMetric(a, now, seen, 'alarm') this.close(ws, RELAY_CLOSE_CODE.PEER_DROPPED, 'control silence timeout') await this.onSocketGone(ws, RELAY_CLOSE_CODE.PEER_DROPPED, '') continue } // Same order as the cells' heartbeat: expired auth past its grace, then the 6 h lease. if (now > a.claims.exp + CONTROL_CONTINUITY_LIMITS.expiredAuthExistingSpliceGraceMs) { + // A host already silent past the limit is gone (asleep, offline): it closes as one, so + // drop counts stay comparable with the cells, whose pings would have closed it earlier. + // Its probe is cut short: the expired authorization ends the socket either way. + if (now - seen > this.config.silenceTimeoutMs) { + this.deadControlMetric(a, now, seen, 'expired') + this.close(ws, RELAY_CLOSE_CODE.PEER_DROPPED, 'control silence timeout') + await this.onSocketGone(ws, RELAY_CLOSE_CODE.PEER_DROPPED, '', Math.min(now, seen + this.config.silenceTimeoutMs + this.config.pingIntervalMs)) + continue + } const sinceRefreshS = a.lastRefreshAt === undefined ? -1 : Math.round((now - a.lastRefreshAt) / 1000) const seenAgoS = Math.round((now - seen) / 1000) - metric(this.env, 'auth_expired_close', sinceRefreshS, a.lastRefreshAt === undefined ? 'never' : 'refreshed', [seenAgoS], this.keepalive(a) ? 'keepalive' : 'legacy', a.place) + metric(this.env, 'auth_expired_close', sinceRefreshS, a.lastRefreshAt === undefined ? 'never' : 'refreshed', [seenAgoS, a.activeAt === undefined ? -1 : Math.round((now - a.activeAt) / 1000)], this.keepalive(a) ? 'keepalive' : 'legacy', a.place) this.close(ws, RELAY_CLOSE_CODE.BAD_OUTER_CREDENTIAL, 'relay authorization expired') await this.onSocketGone(ws, RELAY_CLOSE_CODE.BAD_OUTER_CREDENTIAL, '') continue diff --git a/cloud/apps/relay-edge/src/metrics.ts b/cloud/apps/relay-edge/src/metrics.ts index febc2bab77c..ea76d7f7f2c 100644 --- a/cloud/apps/relay-edge/src/metrics.ts +++ b/cloud/apps/relay-edge/src/metrics.ts @@ -2,7 +2,8 @@ import type { Env } from './env' export type MetricEvent = - // label 'new' | 'rebind'; reason 'keepalive' | 'legacy' (took the keepalive offer or not). + // label 'new' | 'rebind'; reason 'keepalive' | 'legacy' (took the keepalive offer or not); + // double2 = ms from the upgrade to activation (the desktop gives up a connect after 15 s). | 'control_active' | 'control_close' | 'phone_open' @@ -27,8 +28,11 @@ export type MetricEvent = | 'control_upgrade' // Control closed 4401 for expired authorization: value = seconds since its last extending // refresh, label 'refreshed'; -1 and label 'never' when this socket never extended its token. - // double2 = seconds since the host's last inbound frame or keepalive; reason 'keepalive' | 'legacy'. + // double2 = seconds since the host's last inbound frame or keepalive, double3 = seconds since the + // control became active (-1 if unknown); reason 'keepalive' | 'legacy'. | 'auth_expired_close' + // Control closed for silence (deadControlMetric, host-relay-control.ts). + | 'silence_close' // Pre-auth socket refused by the per-source limiter: label 'phone' | 'data' | 'health'. Unpushed-host phones: 'not_on_edge'. | 'preauth_limited' | 'not_on_edge' diff --git a/cloud/apps/relay-edge/test/harness-mvp-object.ts b/cloud/apps/relay-edge/test/harness-mvp-object.ts index 89ec43b1e7c..69530b07307 100644 --- a/cloud/apps/relay-edge/test/harness-mvp-object.ts +++ b/cloud/apps/relay-edge/test/harness-mvp-object.ts @@ -47,14 +47,14 @@ export function newHost(): Host { } } -export async function mint(key: KeyLike, relayHostId: string, issuer = ISSUER): Promise { +export async function mint(key: KeyLike, relayHostId: string, issuer = ISSUER, expiresIn = '5m'): Promise { return new SignJWT({ purpose: 'host-control', prof: `prof-${USER}`, relayHostId }) .setProtectedHeader({ alg: 'ES256', kid: REMOTE ? 'spike' : 'mvp' }) .setSubject(USER) .setIssuer(issuer) .setAudience('orca-relay') .setIssuedAt() - .setExpirationTime('5m') + .setExpirationTime(expiresIn) .sign(key) } diff --git a/cloud/apps/relay-edge/test/mvp-object-hardening.ts b/cloud/apps/relay-edge/test/mvp-object-hardening.ts index 5bd6553acab..1e86d0058ee 100644 --- a/cloud/apps/relay-edge/test/mvp-object-hardening.ts +++ b/cloud/apps/relay-edge/test/mvp-object-hardening.ts @@ -128,6 +128,32 @@ export async function keepaliveProbeCheck(key: KeyLike): Promise { } } +// K4: a keepalive host gone silent when its authorization expires closes as a dropped host (4408), +// not 4401, so edge drop counts stay comparable with the cells (their pings would have closed it). +export async function keepaliveExpiredSilentCheck(key: KeyLike): Promise { + const host = newHost() + await assign(host.relayHostId, 'edge', EPOCH, []) + const started = Date.now() + // Alive for 50 s, then silent: the silence probe (sent ~9 s later) is still pending when the + // 3 s token passes its 60 s grace at ~63 s, so the expiry check is what closes it. + const c = control(host, await mint(key, host.relayHostId, undefined, '3s'), EPOCH, { + caps: `${SHIPPED_CAPS},client-keepalive-v1`, + drop: () => Date.now() - started > 50_000 + }) + await c.client.connect() + try { + const closed = await waitFor(() => c.events.closes.length > 0, 3_000 + 60_000 + 15_000) + const after = Date.now() - started + record( + 'K4 silent keepalive host at token expiry: closed 4408 (not 4401)', + closed && c.events.closes.includes(4408) && !c.events.closes.includes(4401) && after >= 59_000, + `closedAfter=${after}ms pings=${c.events.pings} closes=${c.events.closes}` + ) + } finally { + c.client.closeNow() + } +} + // F5: POST /__admin/drain sends the host's controls back to the director (drain frame, 4503). export async function drainCheck(key: KeyLike): Promise { const host = newHost() diff --git a/cloud/apps/relay-edge/test/mvp-object.ts b/cloud/apps/relay-edge/test/mvp-object.ts index 5558bb2d179..f9a436c1d37 100644 --- a/cloud/apps/relay-edge/test/mvp-object.ts +++ b/cloud/apps/relay-edge/test/mvp-object.ts @@ -41,7 +41,7 @@ import { waitFor, type Host } from './harness-mvp-object' -import { drainCheck, heartbeatCheck, keepaliveCheck, keepaliveProbeCheck, objectHealthCheck } from './mvp-object-hardening' +import { drainCheck, heartbeatCheck, keepaliveCheck, keepaliveExpiredSilentCheck, keepaliveProbeCheck, objectHealthCheck } from './mvp-object-hardening' import { replayCheck, rotationCheck } from './mvp-object-replay' async function mainRemote(): Promise { @@ -107,7 +107,7 @@ async function main(): Promise { await sleep(500) } await run(host, privateKey, () => jwksHits) - await Promise.all([heartbeatCheck(privateKey), keepaliveCheck(privateKey), keepaliveProbeCheck(privateKey), drainCheck(privateKey), replayCheck(), rotationCheck()]) + await Promise.all([heartbeatCheck(privateKey), keepaliveCheck(privateKey), keepaliveProbeCheck(privateKey), keepaliveExpiredSilentCheck(privateKey), drainCheck(privateKey), replayCheck(), rotationCheck()]) await sweepCheck() await preAuthCheck(stateDir) await objectHealthCheck(stateDir) From 48d5bfd76c7df3bc51a91782b507dd660245f672 Mon Sep 17 00:00:00 2001 From: OrcaWin Date: Tue, 10 Nov 2026 16:26:37 -0800 Subject: [PATCH 02/21] test: remove four low-value renderer suites and redundant cases (#40928) Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com> --- config/reliability-gates.jsonc | 66 +++--- .../scripts/ci/sharding/ci-shard-timings.json | 3 - .../agent-terminal-presence.test.ts | 40 ---- .../remote-runtime-pty-batching.test.ts | 173 +-------------- ...connect-multi-pane-materialization.test.ts | 22 -- ...s-close-routing-active-browser-tab.test.ts | 188 ----------------- ...ree-creation-backend-startup-focus.test.ts | 189 ----------------- .../reads/runtime-paged-file-read.test.ts | 51 ----- src/renderer/src/store/repos/repos.test.ts | 125 +---------- .../hosts/browser-remote-tab-creation.test.ts | 197 ------------------ ...-workspace-activation-and-activity.test.ts | 24 +-- 11 files changed, 40 insertions(+), 1038 deletions(-) delete mode 100644 src/renderer/src/components/terminal-pane/provider-notices/agent-presence/agent-terminal-presence.test.ts delete mode 100644 src/renderer/src/hooks/useIpcEvents-close-routing-active-browser-tab.test.ts delete mode 100644 src/renderer/src/lib/workspace/creation/worktree-creation-backend-startup-focus.test.ts delete mode 100644 src/renderer/src/store/slices/browser/hosts/browser-remote-tab-creation.test.ts diff --git a/config/reliability-gates.jsonc b/config/reliability-gates.jsonc index 852bc9e138b..0799cf6b242 100644 --- a/config/reliability-gates.jsonc +++ b/config/reliability-gates.jsonc @@ -4186,10 +4186,10 @@ "providers": ["remote-runtime"], "coveredPlatforms": ["macos"], "coveredProviders": ["remote-runtime"], - "coverageNotes": "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Deterministic main-IPC contract tests cover disconnect-driven close delivery, exactly-once close, per-subscription teardown isolation against a failing socket close and a throwing liveness probe, containment of a throwing renderer send on the unguarded host-close path, and continued suppression of stale payloads from a retired transport. A headed paired-server journey (real Orca host plus a separate paired Orca desktop client) covers hidden-but-mounted reveal, cold-parked reveal, and cold-parked reveal across a disconnect/reconnect. Live Linux and Windows paired-server evidence and real sleep/wake transport loss remain uncollected.", + "coverageNotes": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Deterministic main-IPC contract tests cover disconnect-driven close delivery, exactly-once close, per-subscription teardown isolation against a failing socket close and a throwing liveness probe, containment of a throwing renderer send on the unguarded host-close path, and continued suppression of stale payloads from a retired transport. A headed paired-server journey (real Orca host plus a separate paired Orca desktop client) covers hidden-but-mounted reveal, cold-parked reveal, and cold-parked reveal across a disconnect/reconnect. Live Linux and Windows paired-server evidence and real sleep/wake transport loss remain uncollected.", "motivatingLinks": ["tests/e2e/paired-remote-terminal-parked-reveal-interactivity.spec.ts"], "invariant": "Every renderer-held runtime subscription receives exactly one terminal close event when its transport is retired, including when the retirement advanced the transport generation first, and a single failing teardown never abandons that environment's remaining subscriptions nor escapes into the transport that reported the close. Payload frames from a retired transport stay suppressed. A revealed remote terminal therefore reattaches over a live multiplex connection: its buffer restores, typed input reaches the host PTY, the echo paints without a tab flip, and the PTY converges on the revealed pane grid.", - "oracle": "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): The main IPC contract test subscribes terminal.multiplex through the real handler, disconnects the environment, and asserts the renderer received exactly one {type: close} subscription event. Two isolation tests subscribe a second stream to the same environment and make the first one fail -- in its socket close, and in the liveness probe inside notifyClosed -- then assert the disconnect does not throw, both transports closed, and every close the renderer could still receive was delivered. A third drives a host-initiated close through the transport callback, which is the one notifyClosed call site with no surrounding guard, with a renderer send that throws, and asserts it cannot escape into the WebSocket close handler. A fourth test asserts that after retirement a late response frame is not forwarded and a late transport close does not re-send. The paired-server journey runs three reveal scenarios against one real host and one real paired desktop client, and for each records buffer restore, host-side receipt of the typed marker through an out-of-band host sink file, live paint without a tab flip, and PTY-versus-pane grid convergence.", + "oracle": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): The main IPC contract test subscribes terminal.multiplex through the real handler, disconnects the environment, and asserts the renderer received exactly one {type: close} subscription event. Two isolation tests subscribe a second stream to the same environment and make the first one fail -- in its socket close, and in the liveness probe inside notifyClosed -- then assert the disconnect does not throw, both transports closed, and every close the renderer could still receive was delivered. A third drives a host-initiated close through the transport callback, which is the one notifyClosed call site with no surrounding guard, with a renderer send that throws, and asserts it cannot escape into the WebSocket close handler. A fourth test asserts that after retirement a late response frame is not forwarded and a late transport close does not re-send. The paired-server journey runs three reveal scenarios against one real host and one real paired desktop client, and for each records buffer restore, host-side receipt of the typed marker through an out-of-band host sink file, live paint without a tab flip, and PTY-versus-pane grid convergence.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts", "pnpm exec playwright test tests/e2e/paired-remote-terminal-parked-reveal-interactivity.spec.ts --config tests/playwright.config.ts --project electron-headless --workers=1" @@ -4295,7 +4295,8 @@ "Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims.", "Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical.", "Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed.", - "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed." + "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed.", + "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed." ], "demotionRule": "Demote if a retired runtime transport can leave a renderer subscription without a close event, if a close is delivered more than once, if one failing teardown strands its sibling subscriptions, if payload frames from a retired transport reach the renderer, or if a revealed remote terminal can stay blank or reject input after a reconnect." }, @@ -8520,13 +8521,13 @@ "providers": ["paired-runtime", "ssh"], "coveredPlatforms": ["macos", "linux"], "coveredProviders": [], - "coverageNotes": "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. The direct side-effect-facts selector retains six cases: replay exclusion, live sequence handling, stale disposer ownership, one handoff drain, drain-before-snapshot, and stale replacement-snapshot refusal. Successful current-entry snapshot application is not witnessed by the retained snapshot case. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2162: Reduced inventory-authority source intentionally loses the examples recorded in knownGaps; historical execution is not validation of changed bytes. Historical prior account (preserved verbatim): Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2003: Remove SSH/absent-inventory attach refusal, later-subscriber retry, no-view adoption/takeback, provider-tail/unknown-read projections, failed-keystroke retry and simpler loss reattach examples. Keep the existing-subscriber inventory-ready retry alias only at its retained callback, plus retained held-input cancellation/replacement and overlapping attach/loss fences. The inventory driver also loses the whole empty-worktree-publication registration and folder-host/unverifiable scalar examples. Retire the entire dated10-case attach receipt; retain16 independent receipts only in their original scopes. No retained-memory or real-daemon measurement follows from supplied provider inventories. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1991: Remove initial-snapshot sequence-domain, limited-tail/truncation and unknown-output-sequence aliases whose callbacks are removed. The source retains partial-chunk replay and requested-snapshot retry/admission at supplied snapshot/data seams, not equivalent replacement or measured retained-memory proof. Preserve the existing stream/pending-wait capacity aliases only to their retained owning multiplex callbacks with supplied PTY resolutions. Removing the sibling-PTY heartbeat dedup work-count sample does not establish a new decorative fanout budget; the unrelated-PTY cancellation fence survives. Preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Remove the reasonless-close selector and narrow census aliases to actual retained runtime and injected-inventory callbacks. The removed synthetic census-dedup and isolated pre-boundary timing cases receive no replacement credit; preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1939 (sources1934/1935): Delete the synthetic watcher-synchronization and partial-reconciliation suites. Retire all affected host-fact, capability-owner, activation-residue and retention-cap aliases because the remaining callbacks do not establish those full claims. Keep exact local hook/timer, byte-stream side effects, exit/disposal, scroll-intent cleanup and deterministic local rank/TTL progression callbacks within their mocked-host scope. No physical paired-host retention, resource budget or equivalent replacement evidence is assigned. The generic broad test command remains a runnable discovery command, never a dated execution receipt. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete hook-agent-status projection suite and its two behavioral assertion aliases. Supplied stored hook/title/resume/staleness/rename/question/generation/PTY-retirement/snapshot-version projections are intentionally lost. The separate osc-status-scan gate mentions this source only in existing historical notes and receives no mutation. No measured hidden-output retention, CPU/RSS or resource budget result is claimed by the deleted source or this audit. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Module ownership cleanup retires the modeled cold-park private watcher/render/work-count suites under the current test policy. Their assertions are no longer active evidence. Retained lifecycle tests and headed/headless journeys do not establish equivalent coverage for those removed harness alarms. Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1627 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Retained exact unchanged-source receipts remain the only current execution evidence. Reduced selectors have no fresh execution result. Audit1417 intentionally deletes src/renderer/src/components/terminal-pane/pty-connection-hidden-tui-snapshot-replay.spec.ts: Loses rich hidden snapshot sequence replay/grid resize and forced repaint after hidden snapshot alarms, including live-trigger deduplication. deletes src/renderer/src/components/terminal-pane/terminal-cold-park-verdict-loop.test.tsx: Loses modeled React oscillation alarm; real pane/watchers are not exercised. deletes src/renderer/src/runtime/remote-runtime-terminal-parse-backpressure.test.ts: Loses delayed credit until parsing, malformed frame/unknown stream handling, discarded transformed output recovery and failed ACK rearm alarms. No equivalent replacement coverage or fresh result is claimed. Historical coverage account: Audit1412 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): This audit removes the browser response-after-manual-disconnect example. Its distinct in-flight response fencing alarm is lost; retained queued-call disconnect tests are not an equivalent replacement. Historical commands and results remain historical. Deterministic headed macOS runs launch an isolated Orca desktop server and a separate paired web client. A byte-identical headless run uses an isolated `orca serve` host. Both create six real paired host PTYs with bounded high-output scrollback, prove sustained output while all six client xterms are warm-mounted but hidden causes zero renderer scheduler work, then drive 180 timed real-PTY spinner frames through a second production `session.tabs.subscribeAll` transport and the real Zustand store. Each worktree must publish and apply only its exact initial-working and final-idle titles. Both topologies ordinary-park five xterms without enabling the lossy retention budget, assert bounded cells/heap/timer lag, then restore the bounded authoritative tail exactly once on the original PTY including output produced while parked and continued input/output. The headed oracle additionally keeps every remote target workspace and terminal unmounted in the host renderer before and after client recovery. Production-parity React contracts reproduce and remove the independent cold-park watcher/title feedback loop, including memo-cache discard, StrictMode cleanup replay, detached split-PTY replacement, and the zero-park fast path. A deterministic cross-boundary contract drives 24 real PTY trackers in distinct worktrees through 2,880 decorative spinner and bare-native redraw chunks and counts host publications, serialized bytes, renderer apply calls, and store mutations while preserving exact per-PTY raw chunks and per-worktree semantic transitions. Fake-clock continuation beyond the 30-minute stale boundary requires one deduplicated refresh per worktree every 15 minutes, globally spaced by at least 50 ms, with every mirrored working status still fresh. Exact-identical Pi plus rotating normalized Pi/Grok frames renew coarse working/permission state; rich prompt/tool/question fields retain their original 30-minute lease, and `stateStartedAt`/acknowledgement identity remain stable across title/hook source swaps. A separate headed desktop-client journey relaunches the host main process while retaining the daemon and a daemon-backgrounded PTY, then requires first-observation foreground sync and continued input/output on the same PTY. Unit coverage keeps layout versions out of snapshot output high-waters when a restored snapshot has no output sequence and retries an existing subscriber only after authoritative provider inventory proves the retained PTY exists. Separate headed/headless ACK-starvation tests recover one stalled stream without replacing its PTY. A headed output-drop oracle proves a successful snapshot probe cannot certify a stale live stream when the authoritative PTY sequence advanced beyond the client high-water. Unit coverage requires an unsequenced client to establish a first-probe baseline, remain attached at the same sequence, and recover only after a later probe advances. A capability-disabled run proves legacy hosts retain the prior lossy limit/TTL fallback. Unit tests cover exact-owner capability routing, raw-stream release, singleton side-effect facts with timed handoff cleanup, 64 simultaneous synthetic-title sources with zero decorative client events after convergence, per-client late-subscribe recovery, semantic bell/title transitions, local animation-frame preservation, legacy no-output-pause delivery gating with exact snapshot restore, provider-authoritative snapshots, 128 active streams plus retry after a 129th-stream capacity rejection, capacity-pressure backoff, full split-leaf remint reconciliation, truncation, and manual-disconnect queue fencing. Linux, Windows, live SSH, and production-scale paired hosts remain gaps. The audit removed the E2E parking override/debug-handle suite. Override exposure guards, delay/retention independence, TTL filtering and debug-handle forwarding alarms are lost. Retained real parking journeys do not establish equivalent test-harness configuration coverage. The test-value audit removes separate responsive-stream, intervening-control-frame and capacity-error classification examples. Retained multiplexer checks use fake clocks and captured transport callbacks; their per-stream isolation, deadline and probe-baseline assertions do not prove an actual host restart or equivalent replacement coverage. Historical run commands and results remain historical. The test-value audit removes the standalone retained/live metadata projection and additive-preference negative example from runtime-client-events.test.ts. Its remaining case checks partial-authority event admission through a captured callback; it does not establish full-pair projection, actual peer delivery or equivalent replacement coverage. Historical commands and results remain historical. The isolated accepted-handle-events observer suite is removed. Its runtime/worktree/pane membership, once-only delivery and listener-release ledger alarms are lost. Live hidden-retention evidence does not establish equivalent observer coverage. Audit1292: The entire host-session mirror empty-inventory settle suite is removed. Live-unpublished versus genuinely-empty hydration, same-ID revision response/in-flight fencing and cleared-tracking pending-probe invalidation alarms are lost. The singleton store and inventories were mocked; no real host census or resume delivery or equivalent replacement coverage is claimed. Audit1302: src/main/runtime/session-tabs-inventory-publication.test.ts: Keep actual runtime public inventory authority across renderer epochs/resync, fallback, aborts and invalid graph publishers. Remove copied nonempty publication, private scan counters/collector and known-host replacement, fake refresh that performs the restoration itself, environment capability flag, and private resume-store characterization whose fixture supplies reservation/rollback ordering. Lose isolated refresh/churn/receiver/capability and copied-payload wiring alarms; no equivalent replacement. Reduced selectors require fresh validation; prior full-suite receipts do not validate removed or changed coverage. No equivalent replacement or physical-provider claim. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1318: src/shared/remote-runtime/control/remote-runtime-shared-control-connection.test.ts: Retain real local encrypted WebSocket regressions for pre-ready socket replacement preserving a request, close-before-ready logical subscription cleanup, unrelated RPC surviving another deadline, and pending setup abort leaving a sibling pending. Host server is synthetic and uses production crypto; kept scenarios also retain internal diagnostics/memory assertions unchanged. Remove happy routing, authority-envelope, unknown-ID/log, cleanup enum, binary, keepalive/liveness and production-budget/private socket inventories; intentionally lose those alarms, no physical host authorization, independent crypto or remote provider claim. Reduced selectors need fresh validation; historical full-suite receipts do not validate the changed coverage. No equivalent replacement or physical-provider claim. Audit1346 removes the entire private-state-injected session-tabs rich-status boundary suite. Its stale task/question, neutral-title preservation, owner-change and provider-generation alarms are intentionally lost. Retained real terminal journeys do not establish equivalent replacement coverage. The prior mixed 42-contract execution receipt is historical and must not validate the reduced current selector; fresh validation remains required. Audit1352: focused Linux unit selections validated on source head b638e07d395b76af3ba6b03761a1a0fe7e5e026e. Historical statements above retain their original scope. Intentional test losses are recorded in knownGaps; these runs claim no physical-provider, p95, rendered E2E or cross-platform validation. Audit1360: focused Linux unit selections validated on source head 415c9cffc2d10024131ec4b6721767a22195de6d. Historical statements above retain their original scope. Intentional test losses are recorded in knownGaps; these runs claim no physical-provider, p95, rendered E2E or cross-platform validation. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1391: src/renderer/src/components/terminal-pane/pty-connection-remote-pause-drop-notice.spec.ts reduced. Lose repeated quiet switches avoiding snapshots and legacy host without drop-report capability requesting a snapshot. Retained imported registrar uses fake transport/xterm writes, exercising real gap/restore lifecycle rather than actual terminal rendering. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1396: src/renderer/src/components/terminal-pane/terminal-parked-watcher-reconciliation.test.ts reduced. Retire standalone reconcile set output and positive untouchedFreshSpawn/workspaceOwner copying alarms; no native watcher execution claim. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1402: src/renderer/src/components/terminal-pane/pty-connection-park-reveal-model-snapshot.spec.ts reduced. Loses basic paired serialize/initial replay forwarding check; hidden alt-frame grid/precedence policy remains. No real terminal renderer claim. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1412: src/main/runtime/mobile-session-tabs-agent-status-heartbeat.test.ts reduced. Loses FIFO fairness/spacing, semantic bypass/cancellation, delayed-boundary scheduling and teardown timer-cleanup alarms. Retained fixed emission cases do not replace those outcomes. No equivalent replacement coverage is claimed.", + "coverageNotes": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. The direct side-effect-facts selector retains six cases: replay exclusion, live sequence handling, stale disposer ownership, one handoff drain, drain-before-snapshot, and stale replacement-snapshot refusal. Successful current-entry snapshot application is not witnessed by the retained snapshot case. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2162: Reduced inventory-authority source intentionally loses the examples recorded in knownGaps; historical execution is not validation of changed bytes. Historical prior account (preserved verbatim): Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2003: Remove SSH/absent-inventory attach refusal, later-subscriber retry, no-view adoption/takeback, provider-tail/unknown-read projections, failed-keystroke retry and simpler loss reattach examples. Keep the existing-subscriber inventory-ready retry alias only at its retained callback, plus retained held-input cancellation/replacement and overlapping attach/loss fences. The inventory driver also loses the whole empty-worktree-publication registration and folder-host/unverifiable scalar examples. Retire the entire dated10-case attach receipt; retain16 independent receipts only in their original scopes. No retained-memory or real-daemon measurement follows from supplied provider inventories. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1991: Remove initial-snapshot sequence-domain, limited-tail/truncation and unknown-output-sequence aliases whose callbacks are removed. The source retains partial-chunk replay and requested-snapshot retry/admission at supplied snapshot/data seams, not equivalent replacement or measured retained-memory proof. Preserve the existing stream/pending-wait capacity aliases only to their retained owning multiplex callbacks with supplied PTY resolutions. Removing the sibling-PTY heartbeat dedup work-count sample does not establish a new decorative fanout budget; the unrelated-PTY cancellation fence survives. Preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Remove the reasonless-close selector and narrow census aliases to actual retained runtime and injected-inventory callbacks. The removed synthetic census-dedup and isolated pre-boundary timing cases receive no replacement credit; preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1939 (sources1934/1935): Delete the synthetic watcher-synchronization and partial-reconciliation suites. Retire all affected host-fact, capability-owner, activation-residue and retention-cap aliases because the remaining callbacks do not establish those full claims. Keep exact local hook/timer, byte-stream side effects, exit/disposal, scroll-intent cleanup and deterministic local rank/TTL progression callbacks within their mocked-host scope. No physical paired-host retention, resource budget or equivalent replacement evidence is assigned. The generic broad test command remains a runnable discovery command, never a dated execution receipt. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete hook-agent-status projection suite and its two behavioral assertion aliases. Supplied stored hook/title/resume/staleness/rename/question/generation/PTY-retirement/snapshot-version projections are intentionally lost. The separate osc-status-scan gate mentions this source only in existing historical notes and receives no mutation. No measured hidden-output retention, CPU/RSS or resource budget result is claimed by the deleted source or this audit. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Module ownership cleanup retires the modeled cold-park private watcher/render/work-count suites under the current test policy. Their assertions are no longer active evidence. Retained lifecycle tests and headed/headless journeys do not establish equivalent coverage for those removed harness alarms. Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1627 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Retained exact unchanged-source receipts remain the only current execution evidence. Reduced selectors have no fresh execution result. Audit1417 intentionally deletes src/renderer/src/components/terminal-pane/pty-connection-hidden-tui-snapshot-replay.spec.ts: Loses rich hidden snapshot sequence replay/grid resize and forced repaint after hidden snapshot alarms, including live-trigger deduplication. deletes src/renderer/src/components/terminal-pane/terminal-cold-park-verdict-loop.test.tsx: Loses modeled React oscillation alarm; real pane/watchers are not exercised. deletes src/renderer/src/runtime/remote-runtime-terminal-parse-backpressure.test.ts: Loses delayed credit until parsing, malformed frame/unknown stream handling, discarded transformed output recovery and failed ACK rearm alarms. No equivalent replacement coverage or fresh result is claimed. Historical coverage account: Audit1412 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): This audit removes the browser response-after-manual-disconnect example. Its distinct in-flight response fencing alarm is lost; retained queued-call disconnect tests are not an equivalent replacement. Historical commands and results remain historical. Deterministic headed macOS runs launch an isolated Orca desktop server and a separate paired web client. A byte-identical headless run uses an isolated `orca serve` host. Both create six real paired host PTYs with bounded high-output scrollback, prove sustained output while all six client xterms are warm-mounted but hidden causes zero renderer scheduler work, then drive 180 timed real-PTY spinner frames through a second production `session.tabs.subscribeAll` transport and the real Zustand store. Each worktree must publish and apply only its exact initial-working and final-idle titles. Both topologies ordinary-park five xterms without enabling the lossy retention budget, assert bounded cells/heap/timer lag, then restore the bounded authoritative tail exactly once on the original PTY including output produced while parked and continued input/output. The headed oracle additionally keeps every remote target workspace and terminal unmounted in the host renderer before and after client recovery. Production-parity React contracts reproduce and remove the independent cold-park watcher/title feedback loop, including memo-cache discard, StrictMode cleanup replay, detached split-PTY replacement, and the zero-park fast path. A deterministic cross-boundary contract drives 24 real PTY trackers in distinct worktrees through 2,880 decorative spinner and bare-native redraw chunks and counts host publications, serialized bytes, renderer apply calls, and store mutations while preserving exact per-PTY raw chunks and per-worktree semantic transitions. Fake-clock continuation beyond the 30-minute stale boundary requires one deduplicated refresh per worktree every 15 minutes, globally spaced by at least 50 ms, with every mirrored working status still fresh. Exact-identical Pi plus rotating normalized Pi/Grok frames renew coarse working/permission state; rich prompt/tool/question fields retain their original 30-minute lease, and `stateStartedAt`/acknowledgement identity remain stable across title/hook source swaps. A separate headed desktop-client journey relaunches the host main process while retaining the daemon and a daemon-backgrounded PTY, then requires first-observation foreground sync and continued input/output on the same PTY. Unit coverage keeps layout versions out of snapshot output high-waters when a restored snapshot has no output sequence and retries an existing subscriber only after authoritative provider inventory proves the retained PTY exists. Separate headed/headless ACK-starvation tests recover one stalled stream without replacing its PTY. A headed output-drop oracle proves a successful snapshot probe cannot certify a stale live stream when the authoritative PTY sequence advanced beyond the client high-water. Unit coverage requires an unsequenced client to establish a first-probe baseline, remain attached at the same sequence, and recover only after a later probe advances. A capability-disabled run proves legacy hosts retain the prior lossy limit/TTL fallback. Unit tests cover exact-owner capability routing, raw-stream release, singleton side-effect facts with timed handoff cleanup, 64 simultaneous synthetic-title sources with zero decorative client events after convergence, per-client late-subscribe recovery, semantic bell/title transitions, local animation-frame preservation, legacy no-output-pause delivery gating with exact snapshot restore, provider-authoritative snapshots, 128 active streams plus retry after a 129th-stream capacity rejection, capacity-pressure backoff, full split-leaf remint reconciliation, truncation, and manual-disconnect queue fencing. Linux, Windows, live SSH, and production-scale paired hosts remain gaps. The audit removed the E2E parking override/debug-handle suite. Override exposure guards, delay/retention independence, TTL filtering and debug-handle forwarding alarms are lost. Retained real parking journeys do not establish equivalent test-harness configuration coverage. The test-value audit removes separate responsive-stream, intervening-control-frame and capacity-error classification examples. Retained multiplexer checks use fake clocks and captured transport callbacks; their per-stream isolation, deadline and probe-baseline assertions do not prove an actual host restart or equivalent replacement coverage. Historical run commands and results remain historical. The test-value audit removes the standalone retained/live metadata projection and additive-preference negative example from runtime-client-events.test.ts. Its remaining case checks partial-authority event admission through a captured callback; it does not establish full-pair projection, actual peer delivery or equivalent replacement coverage. Historical commands and results remain historical. The isolated accepted-handle-events observer suite is removed. Its runtime/worktree/pane membership, once-only delivery and listener-release ledger alarms are lost. Live hidden-retention evidence does not establish equivalent observer coverage. Audit1292: The entire host-session mirror empty-inventory settle suite is removed. Live-unpublished versus genuinely-empty hydration, same-ID revision response/in-flight fencing and cleared-tracking pending-probe invalidation alarms are lost. The singleton store and inventories were mocked; no real host census or resume delivery or equivalent replacement coverage is claimed. Audit1302: src/main/runtime/session-tabs-inventory-publication.test.ts: Keep actual runtime public inventory authority across renderer epochs/resync, fallback, aborts and invalid graph publishers. Remove copied nonempty publication, private scan counters/collector and known-host replacement, fake refresh that performs the restoration itself, environment capability flag, and private resume-store characterization whose fixture supplies reservation/rollback ordering. Lose isolated refresh/churn/receiver/capability and copied-payload wiring alarms; no equivalent replacement. Reduced selectors require fresh validation; prior full-suite receipts do not validate removed or changed coverage. No equivalent replacement or physical-provider claim. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1318: src/shared/remote-runtime/control/remote-runtime-shared-control-connection.test.ts: Retain real local encrypted WebSocket regressions for pre-ready socket replacement preserving a request, close-before-ready logical subscription cleanup, unrelated RPC surviving another deadline, and pending setup abort leaving a sibling pending. Host server is synthetic and uses production crypto; kept scenarios also retain internal diagnostics/memory assertions unchanged. Remove happy routing, authority-envelope, unknown-ID/log, cleanup enum, binary, keepalive/liveness and production-budget/private socket inventories; intentionally lose those alarms, no physical host authorization, independent crypto or remote provider claim. Reduced selectors need fresh validation; historical full-suite receipts do not validate the changed coverage. No equivalent replacement or physical-provider claim. Audit1346 removes the entire private-state-injected session-tabs rich-status boundary suite. Its stale task/question, neutral-title preservation, owner-change and provider-generation alarms are intentionally lost. Retained real terminal journeys do not establish equivalent replacement coverage. The prior mixed 42-contract execution receipt is historical and must not validate the reduced current selector; fresh validation remains required. Audit1352: focused Linux unit selections validated on source head b638e07d395b76af3ba6b03761a1a0fe7e5e026e. Historical statements above retain their original scope. Intentional test losses are recorded in knownGaps; these runs claim no physical-provider, p95, rendered E2E or cross-platform validation. Audit1360: focused Linux unit selections validated on source head 415c9cffc2d10024131ec4b6721767a22195de6d. Historical statements above retain their original scope. Intentional test losses are recorded in knownGaps; these runs claim no physical-provider, p95, rendered E2E or cross-platform validation. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1391: src/renderer/src/components/terminal-pane/pty-connection-remote-pause-drop-notice.spec.ts reduced. Lose repeated quiet switches avoiding snapshots and legacy host without drop-report capability requesting a snapshot. Retained imported registrar uses fake transport/xterm writes, exercising real gap/restore lifecycle rather than actual terminal rendering. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1396: src/renderer/src/components/terminal-pane/terminal-parked-watcher-reconciliation.test.ts reduced. Retire standalone reconcile set output and positive untouchedFreshSpawn/workspaceOwner copying alarms; no native watcher execution claim. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1402: src/renderer/src/components/terminal-pane/pty-connection-park-reveal-model-snapshot.spec.ts reduced. Loses basic paired serialize/initial replay forwarding check; hidden alt-frame grid/precedence policy remains. No real terminal renderer claim. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1412: src/main/runtime/mobile-session-tabs-agent-status-heartbeat.test.ts reduced. Loses FIFO fairness/spacing, semantic bypass/cancellation, delayed-boundary scheduling and teardown timer-cleanup alarms. Retained fixed emission cases do not replace those outcomes. No equivalent replacement coverage is claimed.", "motivatingLinks": [ "https://github.com/stablyai/orca/issues/8652", "https://github.com/stablyai/orca/pull/10625" ], "invariant": "A host advertising terminal.paired-parking.v1 keeps the PTY and bounded authoritative history alive while an ordinary hidden-view park destroys the client xterm and releases its raw per-PTY stream. Reveal must restore up to the requested 5,000 rows, parked-time side effects/output, the same PTY identity, and continued input/output. Parked watcher synchronization compares parking semantics rather than fresh object identity, owns setup and cleanup across StrictMode replay, reconciles real PTY/layout changes, and performs no terminal-state scan when nothing is parked. After a host main-process relaunch, unknown local delivery-sync state must not be treated as proof that a surviving daemon PTY is already foregrounded. An empty headed session-tab inventory is authoritative only after the current renderer graph generation publishes a complete inventory and every execution host answers the PTY census; unpublished, resync-pending, transport-failed, and partially scoped inventories remain unverifiable, while genuinely authoritative empty and non-empty inventories settle and a scope-less legacy empty keeps its pre-authority best-effort settle so outdated hosts retain agent auto-resume. The unpublished-empty regression must keep the resume dispatch parked with exactly one daemon process/writer; restoring unconditional empty hydration must release one dispatch and create two live writers. If subscription precedes provider readiness, authoritative inventory must reconsider that existing subscriber without spawning, resizing, or changing reconnect state. A snapshot without an output high-water must keep it unknown rather than borrowing a layout version that can suppress newer live output. Hosts without the capability must gate hidden raw output before xterm scheduling and repaint from the authoritative snapshot on reveal while retaining the existing limit/TTL force-parking fallback. After the first semantic title state, decorative spinner frequency must add zero paired client-event frames and zero full session-tabs work between status-freshness leases regardless of hidden worktree count. Continuous working or permission evidence may refresh each affected worktree once per 15 minutes through one globally spaced FIFO so current and legacy viewers do not decay before 30 minutes; sibling PTYs share that worktree refresh, while local title animation, raw output, and semantic title, status, bell, completion, and query facts remain intact. A paired terminal stream whose delivery credits stop progressing must replace only that stream; command silence first probes authoritative state and replaces the stream if the probe times out or proves the PTY advanced beyond the client's delivered output sequence. When no comparable delivery high-water exists, the first sequenced probe establishes a baseline and only later advancement proves staleness. A same-sequence snapshot remains valid proof of a responsive silent command. A successful status probe may replace a pre-ready shared-control socket without rejecting or duplicating calls already waiting for that transport. Manual disconnect must retain pairing while preventing queued or passive calls and subscriptions from recreating transport until explicit Connect. A paired pane reconnecting to a relaunched host that has not yet published its window graph (an unpublished frame, bare or client-projected) must stay in recovery rather than retire, and the host's first published snapshot carrying the surface must reattach it; a published frame lacking the surface still retires it. The placeholder is only ever sent before the host's graph first publishes; afterwards a worktree with no tabs gets a real empty answer, pushed once to any client that asked during startup, so a paired client opening a never-opened host worktree still gets its first terminal.", - "oracle": "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. The direct side-effect-facts selector retains six cases: replay exclusion, live sequence handling, stale disposer ownership, one handoff drain, drain-before-snapshot, and stale replacement-snapshot refusal. Successful current-entry snapshot application is not witnessed by the retained snapshot case. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2003: Remove SSH/absent-inventory attach refusal, later-subscriber retry, no-view adoption/takeback, provider-tail/unknown-read projections, failed-keystroke retry and simpler loss reattach examples. Keep the existing-subscriber inventory-ready retry alias only at its retained callback, plus retained held-input cancellation/replacement and overlapping attach/loss fences. The inventory driver also loses the whole empty-worktree-publication registration and folder-host/unverifiable scalar examples. Retire the entire dated10-case attach receipt; retain16 independent receipts only in their original scopes. No retained-memory or real-daemon measurement follows from supplied provider inventories. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1991: Remove initial-snapshot sequence-domain, limited-tail/truncation and unknown-output-sequence aliases whose callbacks are removed. The source retains partial-chunk replay and requested-snapshot retry/admission at supplied snapshot/data seams, not equivalent replacement or measured retained-memory proof. Preserve the existing stream/pending-wait capacity aliases only to their retained owning multiplex callbacks with supplied PTY resolutions. Removing the sibling-PTY heartbeat dedup work-count sample does not establish a new decorative fanout budget; the unrelated-PTY cancellation fence survives. Preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Remove the reasonless-close selector and narrow census aliases to actual retained runtime and injected-inventory callbacks. The removed synthetic census-dedup and isolated pre-boundary timing cases receive no replacement credit; preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1939 (sources1934/1935): Delete the synthetic watcher-synchronization and partial-reconciliation suites. Retire all affected host-fact, capability-owner, activation-residue and retention-cap aliases because the remaining callbacks do not establish those full claims. Keep exact local hook/timer, byte-stream side effects, exit/disposal, scroll-intent cleanup and deterministic local rank/TTL progression callbacks within their mocked-host scope. No physical paired-host retention, resource budget or equivalent replacement evidence is assigned. The generic broad test command remains a runnable discovery command, never a dated execution receipt. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete hook-agent-status projection suite and its two behavioral assertion aliases. Supplied stored hook/title/resume/staleness/rename/question/generation/PTY-retirement/snapshot-version projections are intentionally lost. The separate osc-status-scan gate mentions this source only in existing historical notes and receives no mutation. No measured hidden-output retention, CPU/RSS or resource budget result is claimed by the deleted source or this audit. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Retained exact unchanged-source receipts remain the only current execution evidence. Reduced selectors have no fresh execution result. Historical pre-audit oracle account for changed sources: The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Run one byte-identical six-terminal oracle against an isolated headed desktop host and an isolated headless `orca serve` host. Stage at least 1,000,000 xterm cells, hide all six warm managers, emit an initial working title followed by 30 timed real-PTY spinner frames and a final idle title per worktree, and observe a second real session-tabs subscription plus the production renderer store. Require exactly two transport envelopes and two title mutations per worktree, no decorative title crossing either boundary, and positive decoded-envelope bytes. Enable ordinary parking with the lossy retention budget disabled, require exactly one mounted manager and five parked tabs, at most 45% retained cells, no more than 16 MiB heap growth, and under 500 ms timer drift. In headed mode, require the host renderer to remain on its original workspace with zero target terminal managers mounted throughout client park, reveal, and live I/O. While a tab is parked, require authoritative terminal.read to observe new PTY output; reveal it and require the original PTY, a marker within the requested 5,000-row history, the parked marker, and post-reveal input/output. Drive 24 real PTY title trackers in distinct worktrees through 40 decorative spinner frames each, then 40 spinner frames alternating with Cursor's bare native identity redraw, using fake clocks. Require all 2,880 raw chunks exactly and zero host session-tabs publications, serialized bytes, renderer apply calls, and renderer store mutations. Continue decorative evidence beyond 30 minutes, require exactly one refresh per worktree per 15-minute lease through a FIFO spaced by at least 50 ms, and require all 24 viewer statuses to remain working and fresh; then require one exact semantic idle transition and one visible-output chunk per PTY. Require ordinary and interior-Braille title changes to publish. Drive 64 host PTYs through ten 80 ms synthetic title frames, require zero paired client events after the first frame per PTY while all 640 local frames remain observable, attach a late client and require one current frame per PTY, then require semantic bell and idle transitions on both clients. Reconstruct a legacy subscribed stream with no outputPause capability, hide a chatty pane, require no hidden xterm writes, reveal it, and require one authoritative snapshot plus continued live output. Drop and acknowledge output only for one original paired-client stream, prove the fixture process consumed input while the host model advanced and the client stayed stale, then require the command snapshot probe to replace that stream, repaint exact fixture output, preserve authoritative PTY identity and target-tab cardinality, and resume live I/O. For a client without a delivered sequence, require the first numeric probe to establish a baseline without replacement, a same-sequence probe to remain attached, and a later advanced probe to replace only that stream. Withhold the first encrypted shared-control ready frame, start one RPC, trigger a status-probe refresh, and require two connections, one host delivery, successful response, and zero retained request bytes. Admit 128 active streams, reject the 129th as retryable, release one stream, then require the retry to attach and publish a snapshot without multiplying retained subscribers. Disable terminal.paired-parking.v1 and require the same oracle to fail before parking, while the legacy limit-one fallback separately passes. Also preserve truncated first paint, ACK-starved same-PTY recovery, responsive silent-command snapshot probes, dead-stream probe timeout recovery, and queued manual-disconnect fencing.", + "oracle": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. The direct side-effect-facts selector retains six cases: replay exclusion, live sequence handling, stale disposer ownership, one handoff drain, drain-before-snapshot, and stale replacement-snapshot refusal. Successful current-entry snapshot application is not witnessed by the retained snapshot case. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2003: Remove SSH/absent-inventory attach refusal, later-subscriber retry, no-view adoption/takeback, provider-tail/unknown-read projections, failed-keystroke retry and simpler loss reattach examples. Keep the existing-subscriber inventory-ready retry alias only at its retained callback, plus retained held-input cancellation/replacement and overlapping attach/loss fences. The inventory driver also loses the whole empty-worktree-publication registration and folder-host/unverifiable scalar examples. Retire the entire dated10-case attach receipt; retain16 independent receipts only in their original scopes. No retained-memory or real-daemon measurement follows from supplied provider inventories. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1991: Remove initial-snapshot sequence-domain, limited-tail/truncation and unknown-output-sequence aliases whose callbacks are removed. The source retains partial-chunk replay and requested-snapshot retry/admission at supplied snapshot/data seams, not equivalent replacement or measured retained-memory proof. Preserve the existing stream/pending-wait capacity aliases only to their retained owning multiplex callbacks with supplied PTY resolutions. Removing the sibling-PTY heartbeat dedup work-count sample does not establish a new decorative fanout budget; the unrelated-PTY cancellation fence survives. Preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Remove the reasonless-close selector and narrow census aliases to actual retained runtime and injected-inventory callbacks. The removed synthetic census-dedup and isolated pre-boundary timing cases receive no replacement credit; preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1939 (sources1934/1935): Delete the synthetic watcher-synchronization and partial-reconciliation suites. Retire all affected host-fact, capability-owner, activation-residue and retention-cap aliases because the remaining callbacks do not establish those full claims. Keep exact local hook/timer, byte-stream side effects, exit/disposal, scroll-intent cleanup and deterministic local rank/TTL progression callbacks within their mocked-host scope. No physical paired-host retention, resource budget or equivalent replacement evidence is assigned. The generic broad test command remains a runnable discovery command, never a dated execution receipt. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete hook-agent-status projection suite and its two behavioral assertion aliases. Supplied stored hook/title/resume/staleness/rename/question/generation/PTY-retirement/snapshot-version projections are intentionally lost. The separate osc-status-scan gate mentions this source only in existing historical notes and receives no mutation. No measured hidden-output retention, CPU/RSS or resource budget result is claimed by the deleted source or this audit. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Retained exact unchanged-source receipts remain the only current execution evidence. Reduced selectors have no fresh execution result. Historical pre-audit oracle account for changed sources: The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Run one byte-identical six-terminal oracle against an isolated headed desktop host and an isolated headless `orca serve` host. Stage at least 1,000,000 xterm cells, hide all six warm managers, emit an initial working title followed by 30 timed real-PTY spinner frames and a final idle title per worktree, and observe a second real session-tabs subscription plus the production renderer store. Require exactly two transport envelopes and two title mutations per worktree, no decorative title crossing either boundary, and positive decoded-envelope bytes. Enable ordinary parking with the lossy retention budget disabled, require exactly one mounted manager and five parked tabs, at most 45% retained cells, no more than 16 MiB heap growth, and under 500 ms timer drift. In headed mode, require the host renderer to remain on its original workspace with zero target terminal managers mounted throughout client park, reveal, and live I/O. While a tab is parked, require authoritative terminal.read to observe new PTY output; reveal it and require the original PTY, a marker within the requested 5,000-row history, the parked marker, and post-reveal input/output. Drive 24 real PTY title trackers in distinct worktrees through 40 decorative spinner frames each, then 40 spinner frames alternating with Cursor's bare native identity redraw, using fake clocks. Require all 2,880 raw chunks exactly and zero host session-tabs publications, serialized bytes, renderer apply calls, and renderer store mutations. Continue decorative evidence beyond 30 minutes, require exactly one refresh per worktree per 15-minute lease through a FIFO spaced by at least 50 ms, and require all 24 viewer statuses to remain working and fresh; then require one exact semantic idle transition and one visible-output chunk per PTY. Require ordinary and interior-Braille title changes to publish. Drive 64 host PTYs through ten 80 ms synthetic title frames, require zero paired client events after the first frame per PTY while all 640 local frames remain observable, attach a late client and require one current frame per PTY, then require semantic bell and idle transitions on both clients. Reconstruct a legacy subscribed stream with no outputPause capability, hide a chatty pane, require no hidden xterm writes, reveal it, and require one authoritative snapshot plus continued live output. Drop and acknowledge output only for one original paired-client stream, prove the fixture process consumed input while the host model advanced and the client stayed stale, then require the command snapshot probe to replace that stream, repaint exact fixture output, preserve authoritative PTY identity and target-tab cardinality, and resume live I/O. For a client without a delivered sequence, require the first numeric probe to establish a baseline without replacement, a same-sequence probe to remain attached, and a later advanced probe to replace only that stream. Withhold the first encrypted shared-control ready frame, start one RPC, trigger a status-probe refresh, and require two connections, one host delivery, successful response, and zero retained request bytes. Admit 128 active streams, reject the 129th as retryable, release one stream, then require the retry to attach and publish a snapshot without multiplying retained subscribers. Disable terminal.paired-parking.v1 and require the same oracle to fail before parking, while the legacy limit-one fallback separately passes. Also preserve truncated first paint, ACK-starved same-PTY recovery, responsive silent-command snapshot probes, dead-stream probe timeout recovery, and queued manual-disconnect fencing.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-unpublished-host-graph.test.ts", "ORCA_BACKGROUND_LAUNCH=1 pnpm exec playwright test tests/e2e/paired-remote-terminal-host-quit-reconnect-input.spec.ts --config tests/playwright.config.ts --project electron-headless --workers=1", @@ -9017,7 +9018,8 @@ "Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims.", "Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical.", "Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed.", - "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed." + "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed.", + "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed." ], "demotionRule": "Keep experimental or demote to protection none if decorative title frequency changes paired client-event or session-tabs publication volume after convergence, a semantic fact is dropped, paired first paint is blank, passive work reconnects after manual disconnect, a capable hidden paired manager survives ordinary parking, a legacy hidden pane schedules raw xterm output or bypasses its fallback, reveal loses requested bounded history or live PTY identity, retained buffer cells do not fall, or either headed/headless scenario flakes." }, @@ -9196,7 +9198,7 @@ "providers": ["local", "daemon", "ssh", "remote-runtime"], "coveredPlatforms": ["macos"], "coveredProviders": [], - "coverageNotes": "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Module layout SSH policy: Retire the multiplexer private-clock saturation registration, the writer saturation transition-count registration and the changed-source dated writer/backpressure receipt. Retained byte delivery, lane order, fairness, settlement and cancellation cases are unvalidated by that historical receipt. No equivalent replacement, new execution, measured performance, live SSH/WSL or cross-platform proof is claimed. The unchanged WSL protocol receipt retains only its original command scope. Historical prior account (preserved verbatim): Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2070 (sources2068/2069): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/hooks-and-hosted-review.spec'). src/main/ssh/ssh-multiplexer-transport-writer.test.ts: reports saturated epochs and their drain boundary exactly once. src/main/runtime/orca-runtime-tests/worktree-removal-archive-hook-gate.spec.ts: removes and records the waiver when the failure is explicitly overridden; removes without an override record when the hook succeeds; removes when the hook is configured but not requested; removes when no archive hook is configured; holds a desktop removal in the same repo until this one is accepted; joins a desktop removal of the same worktree accepted while this one waited. src/main/runtime/orca-runtime-tests/worktree-ps-agent-row-dismissal.spec.ts: drops the row as soon as the user dismisses it, without waiting for the PTY to exit; tells paired clients to republish on the transition and on the dismissal; rejoins a row through its %s handle after pane ownership clears. src/main/runtime/orca-runtime-tests/hooks-and-hosted-review.spec.ts: reports an error when the SSH filesystem provider is unavailable; reports an SSH orca.yaml as an error without reading any filesystem; reports ok for a local repo hook check; reads SSH issue commands as absent and drops writes without touching any filesystem; allows host integration slug helpers for SSH repos through provider-aware GitHub clients; routes runtime GitHub repo identity helpers through the selected WSL project runtime; routes runtime GitHub issue and work-item actions through the selected WSL project runtime; pins explicit origin preference on runtime open-by-number work item lookups; forwards check-details cancellation without local Git overrides. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. The historical receipt/currentCommand selecting src/main/ssh/ssh-multiplexer-transport-writer.test.ts remains historical and is not a passing result for the reduced current command. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete tail-wait memo suite: its sole current authored callback checks supplied fromTail/blocked field values. Retire all three broader memoization/equivalence/eviction assertion aliases; they are not backed by retained source callbacks. No scan-work budget, sustained output, physical terminal or resource measurement is assigned to this removed test. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1674 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Local macOS evidence covers main-process pending-output and projection-admission caps, saturated-tail wait detection, direct-SSH source retention/accounting, always-on V1 negotiation with legacy peer fallback, exact provider-generation pause ownership, renderer exit/data ordering, decoder input bounds, bounded reconnect scheduling, and WSL stdio transport settlement through deterministic unit contracts. A joined main/runtime oracle exercises renderer-sourced headed semantics and headless-model snapshot semantics without launching a live paired server. Separately, a macOS-hosted Docker OpenSSH run exercises only the deployed Linux relay and direct SSH provider with an exact 256 KiB source-credit plateau, concurrent PTY typing, fixed-size filesystem frames, Git churn, and owner-lease reconnect. Neither deterministic topology labels nor Docker SSH constitute live headed desktop, headless orca serve, or physical WSL evidence; local/daemon, prior-version daemons, Windows named pipes/ConPTY, folder workspaces, and mixed-version clients also remain uncovered. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result.", + "coverageNotes": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Module layout SSH policy: Retire the multiplexer private-clock saturation registration, the writer saturation transition-count registration and the changed-source dated writer/backpressure receipt. Retained byte delivery, lane order, fairness, settlement and cancellation cases are unvalidated by that historical receipt. No equivalent replacement, new execution, measured performance, live SSH/WSL or cross-platform proof is claimed. The unchanged WSL protocol receipt retains only its original command scope. Historical prior account (preserved verbatim): Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2070 (sources2068/2069): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/hooks-and-hosted-review.spec'). src/main/ssh/ssh-multiplexer-transport-writer.test.ts: reports saturated epochs and their drain boundary exactly once. src/main/runtime/orca-runtime-tests/worktree-removal-archive-hook-gate.spec.ts: removes and records the waiver when the failure is explicitly overridden; removes without an override record when the hook succeeds; removes when the hook is configured but not requested; removes when no archive hook is configured; holds a desktop removal in the same repo until this one is accepted; joins a desktop removal of the same worktree accepted while this one waited. src/main/runtime/orca-runtime-tests/worktree-ps-agent-row-dismissal.spec.ts: drops the row as soon as the user dismisses it, without waiting for the PTY to exit; tells paired clients to republish on the transition and on the dismissal; rejoins a row through its %s handle after pane ownership clears. src/main/runtime/orca-runtime-tests/hooks-and-hosted-review.spec.ts: reports an error when the SSH filesystem provider is unavailable; reports an SSH orca.yaml as an error without reading any filesystem; reports ok for a local repo hook check; reads SSH issue commands as absent and drops writes without touching any filesystem; allows host integration slug helpers for SSH repos through provider-aware GitHub clients; routes runtime GitHub repo identity helpers through the selected WSL project runtime; routes runtime GitHub issue and work-item actions through the selected WSL project runtime; pins explicit origin preference on runtime open-by-number work item lookups; forwards check-details cancellation without local Git overrides. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. The historical receipt/currentCommand selecting src/main/ssh/ssh-multiplexer-transport-writer.test.ts remains historical and is not a passing result for the reduced current command. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete tail-wait memo suite: its sole current authored callback checks supplied fromTail/blocked field values. Retire all three broader memoization/equivalence/eviction assertion aliases; they are not backed by retained source callbacks. No scan-work budget, sustained output, physical terminal or resource measurement is assigned to this removed test. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1674 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Local macOS evidence covers main-process pending-output and projection-admission caps, saturated-tail wait detection, direct-SSH source retention/accounting, always-on V1 negotiation with legacy peer fallback, exact provider-generation pause ownership, renderer exit/data ordering, decoder input bounds, bounded reconnect scheduling, and WSL stdio transport settlement through deterministic unit contracts. A joined main/runtime oracle exercises renderer-sourced headed semantics and headless-model snapshot semantics without launching a live paired server. Separately, a macOS-hosted Docker OpenSSH run exercises only the deployed Linux relay and direct SSH provider with an exact 256 KiB source-credit plateau, concurrent PTY typing, fixed-size filesystem frames, Git churn, and owner-lease reconnect. Neither deterministic topology labels nor Docker SSH constitute live headed desktop, headless orca serve, or physical WSL evidence; local/daemon, prior-version daemons, Windows named pipes/ConPTY, folder workspaces, and mixed-version clients also remain uncovered. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result.", "motivatingLinks": [ "https://github.com/stablyai/orca/pull/6836", "https://github.com/stablyai/orca/pull/6858", @@ -9204,7 +9206,7 @@ "https://github.com/stablyai/orca/pull/7054" ], "invariant": "High-volume terminal output must stay bounded across daemon socket writes, SSH relay writer/decoder/source retention/reconnect, main model and projection admission, runtime metadata, detectors, and tail checks, main-to-renderer in-flight bytes, renderer scheduler queues, and hidden-output restore without starving focused input. Every new SSH session offers V1 and every same-build relay supports it; legacy delivery is reachable only when capability negotiation proves the peer cannot use V1. Negotiated SSH spans remain exact and contiguous through activation, exit, cancellation, recovery, desktop projection, and required remote replacement; provisional activation data cannot project, reconnect may release it only into the exact attempt's private recovery quarantine after contract validation, and exit seals that quarantine against later same-token frames until ordered admission or exact cancellation proof. Reconnect checkpoint capture freezes exact provider-generation/PTy admission, cancels queued old work, and waits for the running raw completion before exposing its accepted checkpoint; timeout or failure detaches only that PTy's old model, makes the checkpoint unavailable, and releases the bounded fence without closing the shared provider, while the same failure outside an active exact migration remains generation-fatal and overlapping reconnects preserve any earlier outstanding fence. Pending renderer projection IDs are capped and compacted across split remainders, exit preparation owns its renderer fence through finalization, exit-time cancellation transfers published projections before proof commit, generation close fences late proof while draining exact projection waiters, closed-generation identity compacts without weakening stale rejection, canceled source-delivery retirement retains at most one ordered token per PTY, and pause/resume targets only the exact provider generation. Additional subscribers cannot stall the owner, while the required legacy primary retains backpressure.", - "oracle": "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Module layout SSH policy: Retire the multiplexer private-clock saturation registration, the writer saturation transition-count registration and the changed-source dated writer/backpressure receipt. Retained byte delivery, lane order, fairness, settlement and cancellation cases are unvalidated by that historical receipt. No equivalent replacement, new execution, measured performance, live SSH/WSL or cross-platform proof is claimed. The unchanged WSL protocol receipt retains only its original command scope. Historical prior account (preserved verbatim): Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2070 (sources2068/2069): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/hooks-and-hosted-review.spec'). src/main/ssh/ssh-multiplexer-transport-writer.test.ts: reports saturated epochs and their drain boundary exactly once. src/main/runtime/orca-runtime-tests/worktree-removal-archive-hook-gate.spec.ts: removes and records the waiver when the failure is explicitly overridden; removes without an override record when the hook succeeds; removes when the hook is configured but not requested; removes when no archive hook is configured; holds a desktop removal in the same repo until this one is accepted; joins a desktop removal of the same worktree accepted while this one waited. src/main/runtime/orca-runtime-tests/worktree-ps-agent-row-dismissal.spec.ts: drops the row as soon as the user dismisses it, without waiting for the PTY to exit; tells paired clients to republish on the transition and on the dismissal; rejoins a row through its %s handle after pane ownership clears. src/main/runtime/orca-runtime-tests/hooks-and-hosted-review.spec.ts: reports an error when the SSH filesystem provider is unavailable; reports an SSH orca.yaml as an error without reading any filesystem; reports ok for a local repo hook check; reads SSH issue commands as absent and drops writes without touching any filesystem; allows host integration slug helpers for SSH repos through provider-aware GitHub clients; routes runtime GitHub repo identity helpers through the selected WSL project runtime; routes runtime GitHub issue and work-item actions through the selected WSL project runtime; pins explicit origin preference on runtime open-by-number work item lookups; forwards check-details cancellation without local Git overrides. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. The historical receipt/currentCommand selecting src/main/ssh/ssh-multiplexer-transport-writer.test.ts remains historical and is not a passing result for the reduced current command. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete tail-wait memo suite: its sole current authored callback checks supplied fromTail/blocked field values. Retire all three broader memoization/equivalence/eviction assertion aliases; they are not backed by retained source callbacks. No scan-work budget, sustained output, physical terminal or resource measurement is assigned to this removed test. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1674 retires changed-source assertion and execution scope without equivalent replacement coverage. Existing unchanged-selector receipts remain scoped to their original sources. Historical prior-source oracle: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: The current executable slice injects main-process renderer backlog pressure, then asserts ordinary terminal-wait detection without a joined-tail allocation and with full prompt-family continuity, per-PTY and total pending-output caps, a 1,024-ID projection cap with split-remainder compaction, preserved sequenced-tail metadata, active-pending protection ahead of background trimming, exact provider-generation pause ownership, renderer exit/data ordering, and ACK-gated in-flight bounds. The negotiation seam proves every initial connection and reconnect offers V1, the same-build relay advertises V1 without launch flags, old clients remain token-free, and method-not-found peers fall back without installing source ACK publication. The joined source-intake/multiplex oracle blocks renderer or headless serialization, admits one snapshot-covered span and one trailing span, requires replacement reservation only after the authoritative sequence is known, advances upstream ACK eligibility exactly to the covered boundary after SnapshotEnd, and delivers then ACKs the trailing span through the ordinary live path without cancellation. Source-range ledger and multiplex seams additionally require cumulative byte ACK 40 then 100 for one 100-byte frame to release exact byte credit without early source settlement, admit a contiguous higher-generation recovered token while the prior token remains unsettled, and reject stale stream or source generations without releasing in-flight byte credit. Replacement reservation rolls back every earlier span on a later failure, commit requires the exact frozen transfer states, and an authoritatively reclaimed covered span rejects commit without local trim authority while rollback removes the reservation idempotently. The exit-deadline oracle publishes a source-backed desktop projection, advances a fake clock to cancellation, and requires one projection transfer, proof commit, preparation, and final exit with no provider close, retained obligation, duplicate finalization, or process-lifetime cancellation tombstone; a second controlled proof remains pending across generation close and must never publish final exit, while both paths release the renderer preparation lease exactly once. The model-migration oracle accepts span A, blocks span B's raw emulator completion, begins disconnect migration, and proves no checkpoint or reconnect attach can remain at A; releasing B advances the exact checkpoint to B and projects B once. Its fake-clock arm reaches the 10-second deadline, requires checkpoint-unavailable, one exact model reset, zero retained admission charge and timers, and no effect from late raw settlement. Its two-PTY failure arm rejects one running raw callback under an active exact migration and requires one checkpoint-unavailable result, one model reset, zero provider closes, retained sibling checkpoint and transport liveness, zero charges and timers, and an unaffected other generation; a non-migrating sibling oracle still requires generation close. The intake additionally closes 2,048 sequential provider generations into one exact range, rejects stale events across that range, and preserves an out-of-order live gap until its own close. Direct-SSH contracts additionally prove activation-response settlement before the first source frame, claim-gated projection under same-decoder-turn response/data delivery, exact early-ACK reservation through send settlement or same-token retry, recovery completion visible before lease-held frames transfers those frames only into the private quarantine and admits no output until exact checkpoint-to-recoveryEnd coverage, invalid-checkpoint restore retention through response settlement followed by fresh-token retry and one live source frame, isolated saturated-subscriber eviction with healthy owner liveness, preserved legacy-primary backpressure, one-write ownership across write(false)/drain, control selection ahead of queued ordinary PTY input with FIFO lanes and bounded fairness, liveness rebasing, failed-exit retention through late ACK or exact recovery, response-settlement cancellation authority, token-scoped exit cancellation, exact checkpoint-to-recoveryEnd continuity, stale-owner one-shot retry, no physical PTY teardown on recovery failure, exact provider-generation closure and once-only cleanup when recovery cancellation publication or proof rejects, 16 MiB-plus-1 MiB decoder caps, charged per-PTY/session retention budgets, and eight-wide isolated reattach. The live Electron perf oracle adds hidden-output floods, renderer scheduler queue depth, dropped-output-zero normal scenarios, and active key latency budgets before promotion.", + "oracle": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Module layout SSH policy: Retire the multiplexer private-clock saturation registration, the writer saturation transition-count registration and the changed-source dated writer/backpressure receipt. Retained byte delivery, lane order, fairness, settlement and cancellation cases are unvalidated by that historical receipt. No equivalent replacement, new execution, measured performance, live SSH/WSL or cross-platform proof is claimed. The unchanged WSL protocol receipt retains only its original command scope. Historical prior account (preserved verbatim): Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2070 (sources2068/2069): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/hooks-and-hosted-review.spec'). src/main/ssh/ssh-multiplexer-transport-writer.test.ts: reports saturated epochs and their drain boundary exactly once. src/main/runtime/orca-runtime-tests/worktree-removal-archive-hook-gate.spec.ts: removes and records the waiver when the failure is explicitly overridden; removes without an override record when the hook succeeds; removes when the hook is configured but not requested; removes when no archive hook is configured; holds a desktop removal in the same repo until this one is accepted; joins a desktop removal of the same worktree accepted while this one waited. src/main/runtime/orca-runtime-tests/worktree-ps-agent-row-dismissal.spec.ts: drops the row as soon as the user dismisses it, without waiting for the PTY to exit; tells paired clients to republish on the transition and on the dismissal; rejoins a row through its %s handle after pane ownership clears. src/main/runtime/orca-runtime-tests/hooks-and-hosted-review.spec.ts: reports an error when the SSH filesystem provider is unavailable; reports an SSH orca.yaml as an error without reading any filesystem; reports ok for a local repo hook check; reads SSH issue commands as absent and drops writes without touching any filesystem; allows host integration slug helpers for SSH repos through provider-aware GitHub clients; routes runtime GitHub repo identity helpers through the selected WSL project runtime; routes runtime GitHub issue and work-item actions through the selected WSL project runtime; pins explicit origin preference on runtime open-by-number work item lookups; forwards check-details cancellation without local Git overrides. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. The historical receipt/currentCommand selecting src/main/ssh/ssh-multiplexer-transport-writer.test.ts remains historical and is not a passing result for the reduced current command. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete tail-wait memo suite: its sole current authored callback checks supplied fromTail/blocked field values. Retire all three broader memoization/equivalence/eviction assertion aliases; they are not backed by retained source callbacks. No scan-work budget, sustained output, physical terminal or resource measurement is assigned to this removed test. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1674 retires changed-source assertion and execution scope without equivalent replacement coverage. Existing unchanged-selector receipts remain scoped to their original sources. Historical prior-source oracle: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: The current executable slice injects main-process renderer backlog pressure, then asserts ordinary terminal-wait detection without a joined-tail allocation and with full prompt-family continuity, per-PTY and total pending-output caps, a 1,024-ID projection cap with split-remainder compaction, preserved sequenced-tail metadata, active-pending protection ahead of background trimming, exact provider-generation pause ownership, renderer exit/data ordering, and ACK-gated in-flight bounds. The negotiation seam proves every initial connection and reconnect offers V1, the same-build relay advertises V1 without launch flags, old clients remain token-free, and method-not-found peers fall back without installing source ACK publication. The joined source-intake/multiplex oracle blocks renderer or headless serialization, admits one snapshot-covered span and one trailing span, requires replacement reservation only after the authoritative sequence is known, advances upstream ACK eligibility exactly to the covered boundary after SnapshotEnd, and delivers then ACKs the trailing span through the ordinary live path without cancellation. Source-range ledger and multiplex seams additionally require cumulative byte ACK 40 then 100 for one 100-byte frame to release exact byte credit without early source settlement, admit a contiguous higher-generation recovered token while the prior token remains unsettled, and reject stale stream or source generations without releasing in-flight byte credit. Replacement reservation rolls back every earlier span on a later failure, commit requires the exact frozen transfer states, and an authoritatively reclaimed covered span rejects commit without local trim authority while rollback removes the reservation idempotently. The exit-deadline oracle publishes a source-backed desktop projection, advances a fake clock to cancellation, and requires one projection transfer, proof commit, preparation, and final exit with no provider close, retained obligation, duplicate finalization, or process-lifetime cancellation tombstone; a second controlled proof remains pending across generation close and must never publish final exit, while both paths release the renderer preparation lease exactly once. The model-migration oracle accepts span A, blocks span B's raw emulator completion, begins disconnect migration, and proves no checkpoint or reconnect attach can remain at A; releasing B advances the exact checkpoint to B and projects B once. Its fake-clock arm reaches the 10-second deadline, requires checkpoint-unavailable, one exact model reset, zero retained admission charge and timers, and no effect from late raw settlement. Its two-PTY failure arm rejects one running raw callback under an active exact migration and requires one checkpoint-unavailable result, one model reset, zero provider closes, retained sibling checkpoint and transport liveness, zero charges and timers, and an unaffected other generation; a non-migrating sibling oracle still requires generation close. The intake additionally closes 2,048 sequential provider generations into one exact range, rejects stale events across that range, and preserves an out-of-order live gap until its own close. Direct-SSH contracts additionally prove activation-response settlement before the first source frame, claim-gated projection under same-decoder-turn response/data delivery, exact early-ACK reservation through send settlement or same-token retry, recovery completion visible before lease-held frames transfers those frames only into the private quarantine and admits no output until exact checkpoint-to-recoveryEnd coverage, invalid-checkpoint restore retention through response settlement followed by fresh-token retry and one live source frame, isolated saturated-subscriber eviction with healthy owner liveness, preserved legacy-primary backpressure, one-write ownership across write(false)/drain, control selection ahead of queued ordinary PTY input with FIFO lanes and bounded fairness, liveness rebasing, failed-exit retention through late ACK or exact recovery, response-settlement cancellation authority, token-scoped exit cancellation, exact checkpoint-to-recoveryEnd continuity, stale-owner one-shot retry, no physical PTY teardown on recovery failure, exact provider-generation closure and once-only cleanup when recovery cancellation publication or proof rejects, 16 MiB-plus-1 MiB decoder caps, charged per-PTY/session retention budgets, and eight-wide isolated reattach. The live Electron perf oracle adds hidden-output floods, renderer scheduler queue depth, dropped-output-zero normal scenarios, and active key latency budgets before promotion.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts", "pnpm exec vitest run --config config/vitest.config.ts src/main/runtime/construction/orca-runtime.test.ts", @@ -9367,7 +9369,8 @@ "Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims.", "Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical.", "Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed.", - "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed." + "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed.", + "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed." ], "demotionRule": "Cannot promote without metric artifacts and bounded-byte assertions." }, @@ -9601,9 +9604,9 @@ ], "platforms": ["macos", "linux", "windows"], "providers": ["ssh", "remote-runtime", "wsl"], - "coveredPlatforms": ["macos"], + "coveredPlatforms": ["linux"], "coveredProviders": ["ssh", "remote-runtime"], - "coverageNotes": "Audit3096: remove the deleted reconnect-banner selector and its two UI aliases. The reduced recovery-state selector retains only the deadline cancelling pending backoff and explicit retry owning a new epoch alias; healthy-replacement stale-schedule and independent caller-cutoff cancellation aliases retire. Other retained state cases cover reentrant retry, owner settlement and same-epoch parked revival, without replacing those removed alarms. Prior admission-selector retirement, archived July 28 receipt, unrelated selectors/claims/receipts and old gaps remain unchanged; no fresh execution is claimed. Audit3086: retire the destroyed fresh-session fallback settlement/non-destructive reattach alias and its direct selector; retain the detach/attach selector and assertion unchanged. The retained admission cases cover adopt-before-data and explicitly buffered-death admission without a new spawn, without replacing fallback retirement. July 28 evidence was already archived by audit3077; no duplicate retirement or new execution is claimed. Audit2709 removes src/renderer/src/runtime/web-session-tabs-sync/inventory/remote-runtime-session-tabs-inflight.test.ts: intentionally retire waiting out an older inventory and sharing a fresh post-operation request among concurrent callers. No equivalent replacement or fresh execution is claimed. Dated receipts remain historical evidence for their unchanged selectors only; they do not validate the reduced commands or full current gate. Historical prior account (preserved verbatim): Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/ssh/connection/ssh-connection-manager.test.ts: tears a failed startup down quietly, so its error is not replaced by a disconnect; marks a dial that never reached the host, so setup reports SSH’s reason; removes the registration when bulk teardown fails; leaves a connection whose own reconnect is asking the user alone; clears the pool entry when the closed connection is still the registered one; still reuses a connection that is already up; dials a user's connection test. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2052 (sources2049/2050/2051): Remove complete low-value restored-pane-title and pane-visibility-identity suites and their current selectors. Their supplied snapshot/title and nullable identity-comparison alarms are intentionally lost without replacement. Unrelated dated evidenceRuns remain unchanged and do not validate the reduced commands. No fresh gate execution, actual PTY/provider, rendered, remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Delete the terminal-create idempotency forwarding suite and its prospective selector. Preserve2 unrelated receipts without claiming the lost forwarding alarm was replaced. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete complete transport-unmount and queued-initial-CWD helper suites: despite lifecycle-shaped fixtures, the unmount owner returns a Boolean PTY-id predicate, while CWD cases check nullish field selection and callback counts. Lose their detach/destroy/teardown and one-shot-CWD supplied examples. Partial queued-pane-startup loses three identical/empty/nonowner identity examples, retaining four exact one-shot/replacement ownership callbacks. No assertion object maps to these three sources, so unrelated alias objects remain byte-exact. Preserve the current81ed prior removal of pane-split-startup; stale bd948 source metadata snapshots are not replayed. No actual process/PTY startup, transport shutdown, live SSH or rendered recovery is proven by this metadata change. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Delete the sole supplied public-SSH-state spread/sanitized-error projection and its provider-authority-pair alias. Also delete the complete tiny one-shot split-startup wrapper suite: callback-visible temporary field/reset, sequential payload and thrown-sentinel reset alarms disappear. No SSH provider epoch handoff, connection, process startup or actual terminal split was performed by either deleted suite. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1903 removes whole mocked detach, supplied geometry and/or close-request suites selected by this gate. Their pending-CWD/PTY-binding argument/order, fabricated hit-test and close bookkeeping alarms are intentionally lost without replacement. Valid remaining selectors/assertions and unchanged-selector execution receipts retain their original scope. Reduced unit selections have no fresh broad gate, real PTY/provider, rendered focus, or measured-budget validation. Historical prior account (preserved verbatim): Audit1645 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1342 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Audit1301 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Deterministic renderer coverage proves retained native and runtime SSH payloads are admitted through production routes only with valid complete authority, stale cleanup cannot unregister a replacement runtime terminal, direct SSH Git and folder panes keep their PTY bindings through disconnect and reconnect, one authority chain stops after two automatic attempts even when each timeout exceeds the rolling window, rejected acknowledgements mutate no store maps, and one shared exact attempt admits every concurrent split-pane spawn and reattach while preserving the first PTY as the tab fallback. A later sibling failure rotates the tab once, stale callbacks from the prior attempt mutate no state, split remount activity suppression is counted per leaf, primary PTY exit promotes a bound survivor or preserves an empty continuation gap for a late sibling, and primary, non-primary, or null-PTY detach preserves exact authority on both resulting tabs. Intentional pane disposal cancels its settlement timer without breaking StrictMode remount timeout ownership. A mounted remote-runtime terminal survives repeated transport partitions without changing PTY identity. A real encrypted-WebSocket oracle proves a successful reachability probe can replace a pre-ready shared-control socket without rejecting or duplicating the waiting RPC. Direct SSH coordinator tests cover immediate terminal finalization, hydration correction, damping, bounded retry, and telemetry non-interference. Client/server heartbeat tests cover timer suspension, socket generations fence stale callbacks, cold restored-terminal attachment retries, cached pixels remain unhealthy until authoritative replay, automatic retries stop after one minute, manual reconnect preserves the PTY, and pane closure releases recovery UI state. Current-transport recovery clears every error that transport surfaced without clearing or displaying a sibling pane's errors, while transport suppression, pane retention, and the rendered surface are independently bounded. Current macOS Electron journeys against an ephemeral Linux Docker SSH target cover exact-authority repo/worktree hydration, live terminal recovery after disconnect/reconnect, and eager six-terminal remount after renderer reload. A Windows remote-runtime smoke covers reachability and PTY round-trip. Multi-target live fanout, paired-close, WSL, and patched live partition journeys remain gaps. The test-value audit narrowed resolved-worktree-execution-host.test.ts to cache reuse and invalidation. Its direct hydrated-worktree, missing-folder, restored-owner, host-stamp and empty-runtime selection alarms are lost. Other routing and recovery suites do not establish equivalent selector coverage; historical run commands and results remain historical. The test-value audit removes standalone unwritten/released renderer-authority and cross-scope request-count examples. The retained synthetic mirror/store/sidebar and pending-inventory checks do not establish real PTY byte ownership, SSH reconnection or equivalent replacement coverage; historical commands and results remain historical. The test-value audit removes direct join/label ledgers, immediate failure and negative-only stale/removal/partial-authority examples, plus the standalone retained/live event projection. Retained synthetic store/hydration compositions do not establish real SSH continuity, remote authority delivery or equivalent replacement alarms. Historical commands and results remain historical. The renderer scheduling fixture retains in-flight coalescing only. Its immediate publication and negative stale-SSH-cleanup call ledgers were removed; the retained SSH recovery evidence does not replace those alarms. The accepted-handle-events observer suite is removed, including its scoped listener and settlement-release assertions. These observer alarms are lost; retained SSH/transport journeys are not equivalent replacement coverage. The test-value audit removed repeated remote-handle rotation inventories and snapshot-first pending-restart migration examples from terminal-pty-identity-replacement.test.ts, plus direct foreground metadata setter, confirmation, evidence-enum, and prefix/worktree cleanup examples. Those rotation, pending-restart, snapshot-first migration, and foreground scalar-update alarms are lost. The retained replacement case observes pane and tab identities in one subscriber publication; it does not establish equivalent migration, process-evidence, or input-routing coverage. Historical commands and results remain historical. The test-value audit narrows terminals-hydration.test.ts to a real reconnect batch with one measured store publication and hydration-seeded history consumed by a subsequent visit/Back check. Direct split-layout, runtime/folder/floating restoration, dead-leaf focus, cold/known-absent SSH deferral, title, default-tab marker, navigation initialization, and hydrationSucceeded scalar examples are removed; those specific alarms are lost. No actual target-removal transition, live SSH connection, session writer, or persisted restart is exercised by the retained cases. The old target-scoped/snapshot assertion references are retired because the current file does not contain those oracles. Other suites and historical runs do not establish equivalent replacement coverage. Audit1299 removes tests/e2e/terminal-cold-activation-deferral.spec.ts. Private-state/work-budget/assembled-fixture alarms are intentionally lost; retained suites are not equivalent replacement coverage. No new runtime/platform evidence is claimed. Audit1301: src/renderer/src/runtime/remote-runtime-session-tabs-inflight.test.ts reduced. Direct dedupe eviction/count and equal receipt-position sample loses authored alarm. No equivalent replacement coverage is claimed. Audit1318: src/shared/remote-runtime/control/remote-runtime-shared-control-connection.test.ts: Retain real local encrypted WebSocket regressions for pre-ready socket replacement preserving a request, close-before-ready logical subscription cleanup, unrelated RPC surviving another deadline, and pending setup abort leaving a sibling pending. Host server is synthetic and uses production crypto; kept scenarios also retain internal diagnostics/memory assertions unchanged. Remove happy routing, authority-envelope, unknown-ID/log, cleanup enum, binary, keepalive/liveness and production-budget/private socket inventories; intentionally lose those alarms, no physical host authorization, independent crypto or remote provider claim. Reduced selectors need fresh validation; historical full-suite receipts do not validate the changed coverage. No equivalent replacement or physical-provider claim. Audit1342: src/renderer/src/hooks/ipc-events/direct-ssh-connect-reply-routing.test.ts removed. Initial and event-triggered partial-authority hydration fixture alarms disappear. src/renderer/src/components/terminal-pane/terminal-error-accumulation.test.ts removed. Multiline dedup/storm bounds, resource classification/collision, scoped clear/reconnect strip and1000-pane map count alarms disappear. No equivalent replacement coverage is claimed. Audit1396: src/renderer/src/components/terminal-pane/pty-connection-pending-spawn.spec.ts reduced. Lose standalone debug-state retention/log silence alarm. React StrictMode is not mounted; bindings manually simulate remount over mocked transport. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1400: src/main/ipc/ssh/ssh.test.ts reduced. Lose trivial missing target and mocked publication inventory. Preexisting unresolved ../ssh/ssh-port-scanner mock is recorded, not fixed. src/renderer/src/runtime/sync-runtime-graph-scheduling.test.ts removed. Removes whole coalescing call-count suite; no newest-state delivery or replacement claim. src/renderer/src/store/slices/terminal-pty-identity-replacement.test.ts reduced. Removes three no-op/work-count registrations; retains externally observed ownership consistency. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1412: src/main/ssh/ssh-provider-authority.test.ts removed. Loses real AbortController revocation/target isolation/reentrancy outcomes, recreated-target generation-floor/session rollover behavior, stale mutation refusal and settled-registration cleanup alarms. These useful alarms are deliberately lost; no equivalent cancellation coverage is claimed. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1743 removes the standalone line-height clamp/default and pane cwd fallback helper tests (5 authored cases). These trivial-helper alarms are intentionally lost, with no equivalent replacement. Only current commands/testFiles lose the deleted selectors; historical receipts and evidenceRuns remain unchanged and do not validate the reduced commands.", + "coverageNotes": "Audit3315 execution amendment: the reduced two-suite reconnect selector passed on Linux (2 cases). Original source-review statements about no fresh execution describe the earlier static review; this new receipt validates only these two local deterministic unit cases, not the complete gate or physical/runtime/rendered/provider/platform parity. Historical seven-case macOS receipt remains archived verbatim. Audit 3312 retires the ordinary six-pane distinct-live-handle reconnect/resubscription alarm from paired-reconnect-multi-pane-materialization.test.ts. The unchanged parked-pane activation/rematerialization case remains narrower and is not an equivalent replacement. The complete prior seven-contract macOS receipt is archived as historical evidence; it does not validate the reduced selector. No fresh socket, PTY, sleep/wake, provider, platform, mobile, mixed-version or rendered execution is claimed. Audit3096: remove the deleted reconnect-banner selector and its two UI aliases. The reduced recovery-state selector retains only the deadline cancelling pending backoff and explicit retry owning a new epoch alias; healthy-replacement stale-schedule and independent caller-cutoff cancellation aliases retire. Other retained state cases cover reentrant retry, owner settlement and same-epoch parked revival, without replacing those removed alarms. Prior admission-selector retirement, archived July 28 receipt, unrelated selectors/claims/receipts and old gaps remain unchanged; no fresh execution is claimed. Audit3086: retire the destroyed fresh-session fallback settlement/non-destructive reattach alias and its direct selector; retain the detach/attach selector and assertion unchanged. The retained admission cases cover adopt-before-data and explicitly buffered-death admission without a new spawn, without replacing fallback retirement. July 28 evidence was already archived by audit3077; no duplicate retirement or new execution is claimed. Audit2709 removes src/renderer/src/runtime/web-session-tabs-sync/inventory/remote-runtime-session-tabs-inflight.test.ts: intentionally retire waiting out an older inventory and sharing a fresh post-operation request among concurrent callers. No equivalent replacement or fresh execution is claimed. Dated receipts remain historical evidence for their unchanged selectors only; they do not validate the reduced commands or full current gate. Historical prior account (preserved verbatim): Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/ssh/connection/ssh-connection-manager.test.ts: tears a failed startup down quietly, so its error is not replaced by a disconnect; marks a dial that never reached the host, so setup reports SSH’s reason; removes the registration when bulk teardown fails; leaves a connection whose own reconnect is asking the user alone; clears the pool entry when the closed connection is still the registered one; still reuses a connection that is already up; dials a user's connection test. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2052 (sources2049/2050/2051): Remove complete low-value restored-pane-title and pane-visibility-identity suites and their current selectors. Their supplied snapshot/title and nullable identity-comparison alarms are intentionally lost without replacement. Unrelated dated evidenceRuns remain unchanged and do not validate the reduced commands. No fresh gate execution, actual PTY/provider, rendered, remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Delete the terminal-create idempotency forwarding suite and its prospective selector. Preserve2 unrelated receipts without claiming the lost forwarding alarm was replaced. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete complete transport-unmount and queued-initial-CWD helper suites: despite lifecycle-shaped fixtures, the unmount owner returns a Boolean PTY-id predicate, while CWD cases check nullish field selection and callback counts. Lose their detach/destroy/teardown and one-shot-CWD supplied examples. Partial queued-pane-startup loses three identical/empty/nonowner identity examples, retaining four exact one-shot/replacement ownership callbacks. No assertion object maps to these three sources, so unrelated alias objects remain byte-exact. Preserve the current81ed prior removal of pane-split-startup; stale bd948 source metadata snapshots are not replayed. No actual process/PTY startup, transport shutdown, live SSH or rendered recovery is proven by this metadata change. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Delete the sole supplied public-SSH-state spread/sanitized-error projection and its provider-authority-pair alias. Also delete the complete tiny one-shot split-startup wrapper suite: callback-visible temporary field/reset, sequential payload and thrown-sentinel reset alarms disappear. No SSH provider epoch handoff, connection, process startup or actual terminal split was performed by either deleted suite. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1903 removes whole mocked detach, supplied geometry and/or close-request suites selected by this gate. Their pending-CWD/PTY-binding argument/order, fabricated hit-test and close bookkeeping alarms are intentionally lost without replacement. Valid remaining selectors/assertions and unchanged-selector execution receipts retain their original scope. Reduced unit selections have no fresh broad gate, real PTY/provider, rendered focus, or measured-budget validation. Historical prior account (preserved verbatim): Audit1645 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1342 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Audit1301 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Deterministic renderer coverage proves retained native and runtime SSH payloads are admitted through production routes only with valid complete authority, stale cleanup cannot unregister a replacement runtime terminal, direct SSH Git and folder panes keep their PTY bindings through disconnect and reconnect, one authority chain stops after two automatic attempts even when each timeout exceeds the rolling window, rejected acknowledgements mutate no store maps, and one shared exact attempt admits every concurrent split-pane spawn and reattach while preserving the first PTY as the tab fallback. A later sibling failure rotates the tab once, stale callbacks from the prior attempt mutate no state, split remount activity suppression is counted per leaf, primary PTY exit promotes a bound survivor or preserves an empty continuation gap for a late sibling, and primary, non-primary, or null-PTY detach preserves exact authority on both resulting tabs. Intentional pane disposal cancels its settlement timer without breaking StrictMode remount timeout ownership. A mounted remote-runtime terminal survives repeated transport partitions without changing PTY identity. A real encrypted-WebSocket oracle proves a successful reachability probe can replace a pre-ready shared-control socket without rejecting or duplicating the waiting RPC. Direct SSH coordinator tests cover immediate terminal finalization, hydration correction, damping, bounded retry, and telemetry non-interference. Client/server heartbeat tests cover timer suspension, socket generations fence stale callbacks, cold restored-terminal attachment retries, cached pixels remain unhealthy until authoritative replay, automatic retries stop after one minute, manual reconnect preserves the PTY, and pane closure releases recovery UI state. Current-transport recovery clears every error that transport surfaced without clearing or displaying a sibling pane's errors, while transport suppression, pane retention, and the rendered surface are independently bounded. Current macOS Electron journeys against an ephemeral Linux Docker SSH target cover exact-authority repo/worktree hydration, live terminal recovery after disconnect/reconnect, and eager six-terminal remount after renderer reload. A Windows remote-runtime smoke covers reachability and PTY round-trip. Multi-target live fanout, paired-close, WSL, and patched live partition journeys remain gaps. The test-value audit narrowed resolved-worktree-execution-host.test.ts to cache reuse and invalidation. Its direct hydrated-worktree, missing-folder, restored-owner, host-stamp and empty-runtime selection alarms are lost. Other routing and recovery suites do not establish equivalent selector coverage; historical run commands and results remain historical. The test-value audit removes standalone unwritten/released renderer-authority and cross-scope request-count examples. The retained synthetic mirror/store/sidebar and pending-inventory checks do not establish real PTY byte ownership, SSH reconnection or equivalent replacement coverage; historical commands and results remain historical. The test-value audit removes direct join/label ledgers, immediate failure and negative-only stale/removal/partial-authority examples, plus the standalone retained/live event projection. Retained synthetic store/hydration compositions do not establish real SSH continuity, remote authority delivery or equivalent replacement alarms. Historical commands and results remain historical. The renderer scheduling fixture retains in-flight coalescing only. Its immediate publication and negative stale-SSH-cleanup call ledgers were removed; the retained SSH recovery evidence does not replace those alarms. The accepted-handle-events observer suite is removed, including its scoped listener and settlement-release assertions. These observer alarms are lost; retained SSH/transport journeys are not equivalent replacement coverage. The test-value audit removed repeated remote-handle rotation inventories and snapshot-first pending-restart migration examples from terminal-pty-identity-replacement.test.ts, plus direct foreground metadata setter, confirmation, evidence-enum, and prefix/worktree cleanup examples. Those rotation, pending-restart, snapshot-first migration, and foreground scalar-update alarms are lost. The retained replacement case observes pane and tab identities in one subscriber publication; it does not establish equivalent migration, process-evidence, or input-routing coverage. Historical commands and results remain historical. The test-value audit narrows terminals-hydration.test.ts to a real reconnect batch with one measured store publication and hydration-seeded history consumed by a subsequent visit/Back check. Direct split-layout, runtime/folder/floating restoration, dead-leaf focus, cold/known-absent SSH deferral, title, default-tab marker, navigation initialization, and hydrationSucceeded scalar examples are removed; those specific alarms are lost. No actual target-removal transition, live SSH connection, session writer, or persisted restart is exercised by the retained cases. The old target-scoped/snapshot assertion references are retired because the current file does not contain those oracles. Other suites and historical runs do not establish equivalent replacement coverage. Audit1299 removes tests/e2e/terminal-cold-activation-deferral.spec.ts. Private-state/work-budget/assembled-fixture alarms are intentionally lost; retained suites are not equivalent replacement coverage. No new runtime/platform evidence is claimed. Audit1301: src/renderer/src/runtime/remote-runtime-session-tabs-inflight.test.ts reduced. Direct dedupe eviction/count and equal receipt-position sample loses authored alarm. No equivalent replacement coverage is claimed. Audit1318: src/shared/remote-runtime/control/remote-runtime-shared-control-connection.test.ts: Retain real local encrypted WebSocket regressions for pre-ready socket replacement preserving a request, close-before-ready logical subscription cleanup, unrelated RPC surviving another deadline, and pending setup abort leaving a sibling pending. Host server is synthetic and uses production crypto; kept scenarios also retain internal diagnostics/memory assertions unchanged. Remove happy routing, authority-envelope, unknown-ID/log, cleanup enum, binary, keepalive/liveness and production-budget/private socket inventories; intentionally lose those alarms, no physical host authorization, independent crypto or remote provider claim. Reduced selectors need fresh validation; historical full-suite receipts do not validate the changed coverage. No equivalent replacement or physical-provider claim. Audit1342: src/renderer/src/hooks/ipc-events/direct-ssh-connect-reply-routing.test.ts removed. Initial and event-triggered partial-authority hydration fixture alarms disappear. src/renderer/src/components/terminal-pane/terminal-error-accumulation.test.ts removed. Multiline dedup/storm bounds, resource classification/collision, scoped clear/reconnect strip and1000-pane map count alarms disappear. No equivalent replacement coverage is claimed. Audit1396: src/renderer/src/components/terminal-pane/pty-connection-pending-spawn.spec.ts reduced. Lose standalone debug-state retention/log silence alarm. React StrictMode is not mounted; bindings manually simulate remount over mocked transport. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1400: src/main/ipc/ssh/ssh.test.ts reduced. Lose trivial missing target and mocked publication inventory. Preexisting unresolved ../ssh/ssh-port-scanner mock is recorded, not fixed. src/renderer/src/runtime/sync-runtime-graph-scheduling.test.ts removed. Removes whole coalescing call-count suite; no newest-state delivery or replacement claim. src/renderer/src/store/slices/terminal-pty-identity-replacement.test.ts reduced. Removes three no-op/work-count registrations; retains externally observed ownership consistency. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1412: src/main/ssh/ssh-provider-authority.test.ts removed. Loses real AbortController revocation/target isolation/reentrancy outcomes, recreated-target generation-floor/session rollover behavior, stale mutation refusal and settled-registration cleanup alarms. These useful alarms are deliberately lost; no equivalent cancellation coverage is claimed. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1743 removes the standalone line-height clamp/default and pane cwd fallback helper tests (5 authored cases). These trivial-helper alarms are intentionally lost, with no equivalent replacement. Only current commands/testFiles lose the deleted selectors; historical receipts and evidenceRuns remain unchanged and do not validate the reduced commands.", "motivatingLinks": [ "https://linear.app/stably/issue/STA-3107", "https://github.com/stablyai/orca/pull/12664", @@ -9618,7 +9621,7 @@ "https://github.com/stablyai/orca/issues/12902" ], "invariant": "SSH, WSL, and remote-runtime restore paths must treat provider listing failures and unknown liveness as unknown, not dead, while still avoiding duplicate spawn and clearing expired relay leases exactly once. Direct SSH disconnect, reconnect, and wake must keep PTY bindings and relay identity, and target snapshot hydration/reconnect cannot reset sibling SSH, local, WSL, or runtime-owned state. Every restored remote terminal must preserve its provider PTY identity, including the authoritative incarnation returned by a successful session-ID reattach. After a recoverable partition the same authenticated runtime must reattach the same PTY, reject detached input, apply the latest viewport, and report healthy only after authoritative replay. A successful one-shot reachability probe may replace a pre-ready shared-control socket, but waiting RPCs must continue onto the replacement under their original deadline without duplicate host delivery or retained request bytes. Automatic PTY recovery stops after one bounded minute without a fatal terminal error; a manual reconnect starts a newly fenced epoch against the same PTY, and closed panes retain no recovery UI state. Successful current-transport recovery must clear every stale error surfaced for that pane without clearing or displaying a sibling pane's error; a later genuine failure must remain visible. Each transport retains at most eight suppressed messages, each pane retains at most eight distinct messages, and the rendered surface retains at most 24 lines and 4,000 characters. One capability-gated terminal-create mutation must produce at most one host PTY across an unknown response outcome, remain manually retryable after cutoff, and never let a stale completion replace a newer pane lifecycle. Reconnect must alternate exact activation with authoritative inventory so neither a stale activation response nor an activation failure can strand or retire a pane, and activating a parked surface whose persisted binding was already retired must respawn it rather than report a changed owner after signalling its exit.", - "oracle": "Audit3096: remove the deleted reconnect-banner selector and its two UI aliases. The reduced recovery-state selector retains only the deadline cancelling pending backoff and explicit retry owning a new epoch alias; healthy-replacement stale-schedule and independent caller-cutoff cancellation aliases retire. Other retained state cases cover reentrant retry, owner settlement and same-epoch parked revival, without replacing those removed alarms. Prior admission-selector retirement, archived July 28 receipt, unrelated selectors/claims/receipts and old gaps remain unchanged; no fresh execution is claimed. Audit3086: retire the destroyed fresh-session fallback settlement/non-destructive reattach alias and its direct selector; retain the detach/attach selector and assertion unchanged. The retained admission cases cover adopt-before-data and explicitly buffered-death admission without a new spawn, without replacing fallback retirement. July 28 evidence was already archived by audit3077; no duplicate retirement or new execution is claimed. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/ssh/connection/ssh-connection-manager.test.ts: tears a failed startup down quietly, so its error is not replaced by a disconnect; marks a dial that never reached the host, so setup reports SSH’s reason; removes the registration when bulk teardown fails; leaves a connection whose own reconnect is asking the user alone; clears the pool entry when the closed connection is still the registered one; still reuses a connection that is already up; dials a user's connection test. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2052 (sources2049/2050/2051): Remove complete low-value restored-pane-title and pane-visibility-identity suites and their current selectors. Their supplied snapshot/title and nullable identity-comparison alarms are intentionally lost without replacement. Unrelated dated evidenceRuns remain unchanged and do not validate the reduced commands. No fresh gate execution, actual PTY/provider, rendered, remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Delete the terminal-create idempotency forwarding suite and its prospective selector. Preserve2 unrelated receipts without claiming the lost forwarding alarm was replaced. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete complete transport-unmount and queued-initial-CWD helper suites: despite lifecycle-shaped fixtures, the unmount owner returns a Boolean PTY-id predicate, while CWD cases check nullish field selection and callback counts. Lose their detach/destroy/teardown and one-shot-CWD supplied examples. Partial queued-pane-startup loses three identical/empty/nonowner identity examples, retaining four exact one-shot/replacement ownership callbacks. No assertion object maps to these three sources, so unrelated alias objects remain byte-exact. Preserve the current81ed prior removal of pane-split-startup; stale bd948 source metadata snapshots are not replayed. No actual process/PTY startup, transport shutdown, live SSH or rendered recovery is proven by this metadata change. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Delete the sole supplied public-SSH-state spread/sanitized-error projection and its provider-authority-pair alias. Also delete the complete tiny one-shot split-startup wrapper suite: callback-visible temporary field/reset, sequential payload and thrown-sentinel reset alarms disappear. No SSH provider epoch handoff, connection, process startup or actual terminal split was performed by either deleted suite. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1903 removes whole mocked detach, supplied geometry and/or close-request suites selected by this gate. Their pending-CWD/PTY-binding argument/order, fabricated hit-test and close bookkeeping alarms are intentionally lost without replacement. Valid remaining selectors/assertions and unchanged-selector execution receipts retain their original scope. Reduced unit selections have no fresh broad gate, real PTY/provider, rendered focus, or measured-budget validation. Historical prior account (preserved verbatim): Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): Deterministic tests cover bounded stale-handle replacement, suspended heartbeat clocks, cold and established subscription failure, ten partition/recovery cycles, automatic-recovery cutoff, manual reconnect, and direct SSH binding retention. They assert SSH state changes leave PTY bindings in place, relay identity survives, and another target/local/WSL/runtime panes remain byte-identical through target snapshot hydration and reconnect. Both remount callbacks consume split-count activity suppression, intentional dispose emits no failure/timeout, and a same-attempt StrictMode remount still owns one timeout. Hydration clears an untrusted PTY hint without clearing its current pending owner, healthy current-authority bindings suppress correction, hydration finalizes once, and reconnect emits no paired close lifecycle. A provider-level session-ID reattach returns an incarnation, then a legacy exit without an incarnation must resolve to that returned identity rather than minting a fallback identity. The shared-control oracle withholds the first encrypted ready frame, starts one RPC, triggers the successful-probe refresh, then requires exactly two client connections, one host request, a successful response, zero pending calls, and zero retained request bytes. Tests also assert one unsubscribe per remote-runtime epoch, observable recovery phases, stable PTY identity, resumed snapshot/output/input, no healthy state before replay, no retry or input after cutoff, a new manual epoch against the same PTY, quiet recovery UI with an explicit Reconnect action, pane-close state cleanup, one stable create mutation id, old-runtime no-retry behavior, cross-process PTY adoption, and bounded in-flight coordination. A focused recovery oracle surfaces two distinct failures through one transport, completes authoritative replay, and requires both matching clear callbacks; pane-state tests require only the active pane's error to render, matching-pane recovery to preserve sibling and unrelated errors, dismissal to re-admit later failures, and distinct storms to retain only the newest eight messages within 24 lines and 4,000 characters. The retained store replacement assertion observes one atomic pane/tab subscriber publication; repeated rotations and snapshot-first pending-restart migration no longer have their direct store examples.", + "oracle": "Audit 3312 retires the ordinary six-pane distinct-live-handle reconnect/resubscription alarm from paired-reconnect-multi-pane-materialization.test.ts. The unchanged parked-pane activation/rematerialization case remains narrower and is not an equivalent replacement. The complete prior seven-contract macOS receipt is archived as historical evidence; it does not validate the reduced selector. No fresh socket, PTY, sleep/wake, provider, platform, mobile, mixed-version or rendered execution is claimed. Audit3096: remove the deleted reconnect-banner selector and its two UI aliases. The reduced recovery-state selector retains only the deadline cancelling pending backoff and explicit retry owning a new epoch alias; healthy-replacement stale-schedule and independent caller-cutoff cancellation aliases retire. Other retained state cases cover reentrant retry, owner settlement and same-epoch parked revival, without replacing those removed alarms. Prior admission-selector retirement, archived July 28 receipt, unrelated selectors/claims/receipts and old gaps remain unchanged; no fresh execution is claimed. Audit3086: retire the destroyed fresh-session fallback settlement/non-destructive reattach alias and its direct selector; retain the detach/attach selector and assertion unchanged. The retained admission cases cover adopt-before-data and explicitly buffered-death admission without a new spawn, without replacing fallback retirement. July 28 evidence was already archived by audit3077; no duplicate retirement or new execution is claimed. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/ssh/connection/ssh-connection-manager.test.ts: tears a failed startup down quietly, so its error is not replaced by a disconnect; marks a dial that never reached the host, so setup reports SSH’s reason; removes the registration when bulk teardown fails; leaves a connection whose own reconnect is asking the user alone; clears the pool entry when the closed connection is still the registered one; still reuses a connection that is already up; dials a user's connection test. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2052 (sources2049/2050/2051): Remove complete low-value restored-pane-title and pane-visibility-identity suites and their current selectors. Their supplied snapshot/title and nullable identity-comparison alarms are intentionally lost without replacement. Unrelated dated evidenceRuns remain unchanged and do not validate the reduced commands. No fresh gate execution, actual PTY/provider, rendered, remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Delete the terminal-create idempotency forwarding suite and its prospective selector. Preserve2 unrelated receipts without claiming the lost forwarding alarm was replaced. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete complete transport-unmount and queued-initial-CWD helper suites: despite lifecycle-shaped fixtures, the unmount owner returns a Boolean PTY-id predicate, while CWD cases check nullish field selection and callback counts. Lose their detach/destroy/teardown and one-shot-CWD supplied examples. Partial queued-pane-startup loses three identical/empty/nonowner identity examples, retaining four exact one-shot/replacement ownership callbacks. No assertion object maps to these three sources, so unrelated alias objects remain byte-exact. Preserve the current81ed prior removal of pane-split-startup; stale bd948 source metadata snapshots are not replayed. No actual process/PTY startup, transport shutdown, live SSH or rendered recovery is proven by this metadata change. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Delete the sole supplied public-SSH-state spread/sanitized-error projection and its provider-authority-pair alias. Also delete the complete tiny one-shot split-startup wrapper suite: callback-visible temporary field/reset, sequential payload and thrown-sentinel reset alarms disappear. No SSH provider epoch handoff, connection, process startup or actual terminal split was performed by either deleted suite. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1903 removes whole mocked detach, supplied geometry and/or close-request suites selected by this gate. Their pending-CWD/PTY-binding argument/order, fabricated hit-test and close bookkeeping alarms are intentionally lost without replacement. Valid remaining selectors/assertions and unchanged-selector execution receipts retain their original scope. Reduced unit selections have no fresh broad gate, real PTY/provider, rendered focus, or measured-budget validation. Historical prior account (preserved verbatim): Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): Deterministic tests cover bounded stale-handle replacement, suspended heartbeat clocks, cold and established subscription failure, ten partition/recovery cycles, automatic-recovery cutoff, manual reconnect, and direct SSH binding retention. They assert SSH state changes leave PTY bindings in place, relay identity survives, and another target/local/WSL/runtime panes remain byte-identical through target snapshot hydration and reconnect. Both remount callbacks consume split-count activity suppression, intentional dispose emits no failure/timeout, and a same-attempt StrictMode remount still owns one timeout. Hydration clears an untrusted PTY hint without clearing its current pending owner, healthy current-authority bindings suppress correction, hydration finalizes once, and reconnect emits no paired close lifecycle. A provider-level session-ID reattach returns an incarnation, then a legacy exit without an incarnation must resolve to that returned identity rather than minting a fallback identity. The shared-control oracle withholds the first encrypted ready frame, starts one RPC, triggers the successful-probe refresh, then requires exactly two client connections, one host request, a successful response, zero pending calls, and zero retained request bytes. Tests also assert one unsubscribe per remote-runtime epoch, observable recovery phases, stable PTY identity, resumed snapshot/output/input, no healthy state before replay, no retry or input after cutoff, a new manual epoch against the same PTY, quiet recovery UI with an explicit Reconnect action, pane-close state cleanup, one stable create mutation id, old-runtime no-retry behavior, cross-process PTY adoption, and bounded in-flight coordination. A focused recovery oracle surfaces two distinct failures through one transport, completes authoritative replay, and requires both matching clear callbacks; pane-state tests require only the active pane's error to render, matching-pane recovery to preserve sibling and unrelated errors, dismissal to re-admit later failures, and distinct storms to retain only the newest eight messages within 24 lines and 4,000 characters. The retained store replacement assertion observes one atomic pane/tab subscriber publication; repeated rotations and snapshot-first pending-restart migration no longer have their direct store examples.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts src/renderer/src/runtime/web-session-tabs-sync/agent-status/paired-reconnect-sidebar-agent-count.test.ts --reporter=dot", "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal/background-terminal-worktree-mount.test.ts src/renderer/src/components/terminal-pane/mounting/queued-pane-startup.test.ts src/renderer/src/components/terminal-pane/parking/one-shot-park-replay.test.ts src/renderer/src/components/terminal-pane/pty-connection/tests/pty-connection-spawn-and-startup.test.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-host-surface-replacement.test.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-stream-reconnect.test.ts src/renderer/src/store/terminals/terminal-pty-identity-replacement.test.ts", @@ -9667,7 +9670,6 @@ { "file": "src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts", "assertions": [ - "six panes sharing one dropped multiplex connection each resubscribe to a distinct live host handle", "a pane parked while the client slept is authoritatively activated and rematerialized without leaving any pane handle null" ] }, @@ -9791,14 +9793,14 @@ ], "evidenceRuns": [ { - "date": "2026-08-23", + "date": "2026-11-11", "runner": "local", - "platform": "macos", - "command": "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/paired-reconnect-multi-pane-materialization.test.ts src/renderer/src/runtime/paired-reconnect-sidebar-agent-count.test.ts --reporter=dot", + "platform": "linux", + "command": "ORCA_BACKGROUND_LAUNCH=1 pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts src/renderer/src/runtime/web-session-tabs-sync/agent-status/paired-reconnect-sidebar-agent-count.test.ts --reporter=dot --maxWorkers=2", + "currentCommand": "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts src/renderer/src/runtime/web-session-tabs-sync/agent-status/paired-reconnect-sidebar-agent-count.test.ts --reporter=dot", "result": "passed", - "durationSeconds": 2.54, - "summary": "Seven deterministic STA-3107 contracts passed across six-pane multiplex rematerialization and client-owned sidebar-row retention after sleep/wake.", - "currentCommand": "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts src/renderer/src/runtime/web-session-tabs-sync/agent-status/paired-reconnect-sidebar-agent-count.test.ts --reporter=dot" + "durationSeconds": 5.65, + "summary": "Audit3315 focused candidate run: 2 suites / 2 deterministic unit cases passed for parked-pane reconnect materialization and client-owned sidebar retention. currentCommand names the same gate selector; executed command adds background launch and two-worker flags. This is local fake-runtime/store evidence only, not the complete gate, authenticated SSH/PTY/network, rendered Electron, provider, sleep/wake, mobile, mixed-version or cross-platform validation. Re-executed after rebasing onto main 1af64194a46d338777058c06954c2c8ee4a437be (candidate source commit 63adee8174d388466f9276a893972255cee06588)." } ], "runtimeBudget": { @@ -9877,7 +9879,10 @@ "Audit3086: retire the destroyed fresh-session fallback settlement/non-destructive reattach alias and its direct selector; retain the detach/attach selector and assertion unchanged. The retained admission cases cover adopt-before-data and explicitly buffered-death admission without a new spawn, without replacing fallback retirement. July 28 evidence was already archived by audit3077; no duplicate retirement or new execution is claimed.", "Audit3096: remove the deleted reconnect-banner selector and its two UI aliases. The reduced recovery-state selector retains only the deadline cancelling pending backoff and explicit retry owning a new epoch alias; healthy-replacement stale-schedule and independent caller-cutoff cancellation aliases retire. Other retained state cases cover reentrant retry, owner settlement and same-epoch parked revival, without replacing those removed alarms. Prior admission-selector retirement, archived July 28 receipt, unrelated selectors/claims/receipts and old gaps remain unchanged; no fresh execution is claimed.", "Module-layout wave1 intentionally retires complete prohibited owning registrations in src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-attach-subscription.test.ts. Lost alarms: attaches to an existing remote runtime terminal handle; reports a rejected multiplex write through the pane recovery callback; does not report a rejected write from a superseded multiplex stream; does not report attachment health until the authoritative PTY snapshot arrives; drops its side-effect gauge from the census on %s; recovers when the first restored-terminal subscription attempt is offline; recovers when the runtime closes before a restored subscription becomes ready; surfaces a fatal error during subscription setup exactly once; reports viewer disconnect as an unverifiable remote exit; clears every retained transport error when the current stream becomes healthy. No equivalent replacement coverage or pass for a reduced selector is claimed. All unrelated selectors and historical receipts retain their original scope.", - "Module-layout wave2 intentionally retires complete prohibited owning registrations in src/renderer/src/lib/resolved-worktree-execution-host.test.ts. Lost alarms: Both explicit and legacy workspace owners resolve to env-a initially. Replacing the repo catalog refreshes legacy ownership to env-b. A conflicting local repo row makes ownership ambiguous and returns null. Repeated unrelated updates perform zero id reads, an internal work-budget alarm removed. No equivalent replacement coverage or reduced-selector pass is claimed." + "Module-layout wave2 intentionally retires complete prohibited owning registrations in src/renderer/src/lib/resolved-worktree-execution-host.test.ts. Lost alarms: Both explicit and legacy workspace owners resolve to env-a initially. Replacing the repo catalog refreshes legacy ownership to env-b. A conflicting local repo row makes ownership ambiguous and returns null. Repeated unrelated updates perform zero id reads, an internal work-budget alarm removed. No equivalent replacement coverage or reduced-selector pass is claimed.", + "Audit 3312 retires the ordinary six-pane distinct-live-handle reconnect/resubscription alarm from paired-reconnect-multi-pane-materialization.test.ts. The unchanged parked-pane activation/rematerialization case remains narrower and is not an equivalent replacement. The complete prior seven-contract macOS receipt is archived as historical evidence; it does not validate the reduced selector. No fresh socket, PTY, sleep/wake, provider, platform, mobile, mixed-version or rendered execution is claimed.", + "Audit 3312 retires this complete dated seven-contract receipt because paired-reconnect-multi-pane-materialization.test.ts was reduced. Original command, currentCommand, result, counts, date and duration remain historical and do not validate the reduced selector. Original receipt verbatim:\n{\n \"date\": \"2026-08-23\",\n \"runner\": \"local\",\n \"platform\": \"macos\",\n \"command\": \"pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/paired-reconnect-multi-pane-materialization.test.ts src/renderer/src/runtime/paired-reconnect-sidebar-agent-count.test.ts --reporter=dot\",\n \"result\": \"passed\",\n \"durationSeconds\": 2.54,\n \"summary\": \"Seven deterministic STA-3107 contracts passed across six-pane multiplex rematerialization and client-owned sidebar-row retention after sleep/wake.\",\n \"currentCommand\": \"pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts src/renderer/src/runtime/web-session-tabs-sync/agent-status/paired-reconnect-sidebar-agent-count.test.ts --reporter=dot\"\n }", + "Audit3315 execution amendment: the reduced two-suite reconnect selector passed on Linux (2 cases). Original source-review statements about no fresh execution describe the earlier static review; this new receipt validates only these two local deterministic unit cases, not the complete gate or physical/runtime/rendered/provider/platform parity. Historical seven-case macOS receipt remains archived verbatim." ], "demotionRule": "Cannot promote if provider failure can close panes or if the oracle is screenshot-only." }, @@ -11830,14 +11835,14 @@ "providers": ["local", "daemon", "ssh"], "coveredPlatforms": ["macos"], "coveredProviders": ["local"], - "coverageNotes": "Audit 3292 retires the empty-queue immediate query-reply delivery example from remote-runtime-pty-query-reply-immediate.test.ts. Pending ordinary-input separation and asynchronous paste-validation ordering cases remain byte-identical, matching the existing assertion reference. The no-pending-input delivery alarm is intentionally lost without equivalent replacement. Existing selectors, unchanged local Electron receipts and historical receipts remain exact; no new runtime, PTY deadline, platform or mixed-version result is claimed. Audit3086: retire the immediate-reply ceiling and thrown renderer-write recovery aliases. Narrow FIFO coverage to deferred preconnect validation and teardown coverage to natural-exit stale ordinary/accepted tails; live acknowledged barriers and disconnect/detach branches are intentionally retired. Prior queue cleanup and unchanged Electron receipts remain exact; no replacement or fresh execution is claimed. Audit1878 retires the dated reduced-source unit receipt. Only the unchanged macOS local Electron receipts remain current, scoped to their exact commands; reduced unit selectors and SSH have no current execution receipt. Historical pre-audit coverageNotes: Deterministic renderer and IPC-transport tests prove count and text ceilings, oldest-reply shedding, explicit query-reply source routing, ordinary-input preservation, acknowledged-write FIFO barriers, single-worker drain reentrancy, one-reply-per-write delivery for OSC, DA1, and CPR replies, real xterm OSC reply generation, drain-failure containment, teardown settlement, and clear/reuse generation fencing. Remote-runtime tests preserve separate query-reply writes across pending input, async validation, and viewport-claim buffering. Host-contract tests prove a later DA1/CPR reply cannot overtake a deferred OSC reply, including a coalesced legacy-client payload. Live macOS Electron tests cover local PTY OSC replies and interactive typing; a macOS-hosted Docker OpenSSH test proves an upstream-node-pty Linux relay keeps OSC/DA1 replies out of the next fish child's stdin. No live daemon, paired-runtime, WSL, physical Linux/Windows client, or binary mixed-version run is registered.", + "coverageNotes": "Audit 3312 reduces remote-runtime-pty-batching.test.ts, intentionally losing idle leading/trailing coalescing, leading-edge-disabled, accumulated byte-ceiling flush, standalone synchronous oversized-input rejection, accepted-input validation yield, deferred pending-tail byte order, and queued-viewport clear/flush and timer-cancellation alarms. The unchanged UTF-8 splitting, validation barrier ordering and asynchronous oversized-input nondelivery cases are narrower, not equivalent replacements. The barrier assertion and both existing Electron receipts remain unchanged and do not validate these retired cases. No fresh runtime, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3292 retires the empty-queue immediate query-reply delivery example from remote-runtime-pty-query-reply-immediate.test.ts. Pending ordinary-input separation and asynchronous paste-validation ordering cases remain byte-identical, matching the existing assertion reference. The no-pending-input delivery alarm is intentionally lost without equivalent replacement. Existing selectors, unchanged local Electron receipts and historical receipts remain exact; no new runtime, PTY deadline, platform or mixed-version result is claimed. Audit3086: retire the immediate-reply ceiling and thrown renderer-write recovery aliases. Narrow FIFO coverage to deferred preconnect validation and teardown coverage to natural-exit stale ordinary/accepted tails; live acknowledged barriers and disconnect/detach branches are intentionally retired. Prior queue cleanup and unchanged Electron receipts remain exact; no replacement or fresh execution is claimed. Audit1878 retires the dated reduced-source unit receipt. Only the unchanged macOS local Electron receipts remain current, scoped to their exact commands; reduced unit selectors and SSH have no current execution receipt. Historical pre-audit coverageNotes: Deterministic renderer and IPC-transport tests prove count and text ceilings, oldest-reply shedding, explicit query-reply source routing, ordinary-input preservation, acknowledged-write FIFO barriers, single-worker drain reentrancy, one-reply-per-write delivery for OSC, DA1, and CPR replies, real xterm OSC reply generation, drain-failure containment, teardown settlement, and clear/reuse generation fencing. Remote-runtime tests preserve separate query-reply writes across pending input, async validation, and viewport-claim buffering. Host-contract tests prove a later DA1/CPR reply cannot overtake a deferred OSC reply, including a coalesced legacy-client payload. Live macOS Electron tests cover local PTY OSC replies and interactive typing; a macOS-hosted Docker OpenSSH test proves an upstream-node-pty Linux relay keeps OSC/DA1 replies out of the next fish child's stdin. No live daemon, paired-runtime, WSL, physical Linux/Windows client, or binary mixed-version run is registered.", "motivatingLinks": [ "https://github.com/stablyai/orca/issues/13137", "https://github.com/stablyai/orca/issues/7329", "https://github.com/stablyai/orca/issues/13892" ], "invariant": "The desktop PTY input queue retains at most 64 explicitly sourced pending terminal query replies and 4096 UTF-16 code units. Every retained reply reaches the provider as one atomic write, and ordinary, acknowledged, and reply input share one invocation-ordered FIFO so no later write overtakes an acknowledged write. Reentrant write callbacks cannot start a second drain worker or strand input admitted after clear/reuse. A reply's own echo is contained on the output side by projecting its known echo shapes; the ESC-initial verbatim shape is matched only when complete, never held as a partial, so a query torn at its own ESC is still answered. Overflow removes only the oldest query replies, never ordinary input except the documented modified-F3/CPR byte collision, and failure or teardown cannot clear a newer queue generation or strand an acknowledged promise.", - "oracle": "Audit 3292 retires the empty-queue immediate query-reply delivery example from remote-runtime-pty-query-reply-immediate.test.ts. Pending ordinary-input separation and asynchronous paste-validation ordering cases remain byte-identical, matching the existing assertion reference. The no-pending-input delivery alarm is intentionally lost without equivalent replacement. Existing selectors, unchanged local Electron receipts and historical receipts remain exact; no new runtime, PTY deadline, platform or mixed-version result is claimed. Audit3086: retire the immediate-reply ceiling and thrown renderer-write recovery aliases. Narrow FIFO coverage to deferred preconnect validation and teardown coverage to natural-exit stale ordinary/accepted tails; live acknowledged barriers and disconnect/detach branches are intentionally retired. Prior queue cleanup and unchanged Electron receipts remain exact; no replacement or fresh execution is claimed. Audit1878 retires the dated reduced-source unit receipt. Only the unchanged macOS local Electron receipts remain current, scoped to their exact commands; reduced unit selectors and SSH have no current execution receipt. Historical pre-audit oracle: Synchronously enqueue separate 10,000-entry OSC and DA1 reply floods before the scheduled drain and assert that only the initial immediate reply and newest 64 pending replies are written, each as one provider write, before a trailing keystroke. Stall an acknowledged write between earlier and later ordinary/reply input, then require teardown to settle it and same-id reuse to receive no stale tail. Reenter the queue synchronously from an acknowledged write with both enqueue and clear/reuse, requiring one drain and fresh input delivery only after the stale acknowledged write settles false. At the host boundary, defer an OSC reply and assert that separate or legacy-coalesced DA1/CPR replies flush after it in observed query order. At the remote-runtime boundary, preserve separate writes around pending ordinary input, async validation, and viewport-claim buffering. Repeat behind 10,000 ordinary inputs and exercise the text ceiling, real xterm generation, provider-write failure, rejected yield, and clear/reuse generation fencing.", + "oracle": "Audit 3312 reduces remote-runtime-pty-batching.test.ts, intentionally losing idle leading/trailing coalescing, leading-edge-disabled, accumulated byte-ceiling flush, standalone synchronous oversized-input rejection, accepted-input validation yield, deferred pending-tail byte order, and queued-viewport clear/flush and timer-cancellation alarms. The unchanged UTF-8 splitting, validation barrier ordering and asynchronous oversized-input nondelivery cases are narrower, not equivalent replacements. The barrier assertion and both existing Electron receipts remain unchanged and do not validate these retired cases. No fresh runtime, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3292 retires the empty-queue immediate query-reply delivery example from remote-runtime-pty-query-reply-immediate.test.ts. Pending ordinary-input separation and asynchronous paste-validation ordering cases remain byte-identical, matching the existing assertion reference. The no-pending-input delivery alarm is intentionally lost without equivalent replacement. Existing selectors, unchanged local Electron receipts and historical receipts remain exact; no new runtime, PTY deadline, platform or mixed-version result is claimed. Audit3086: retire the immediate-reply ceiling and thrown renderer-write recovery aliases. Narrow FIFO coverage to deferred preconnect validation and teardown coverage to natural-exit stale ordinary/accepted tails; live acknowledged barriers and disconnect/detach branches are intentionally retired. Prior queue cleanup and unchanged Electron receipts remain exact; no replacement or fresh execution is claimed. Audit1878 retires the dated reduced-source unit receipt. Only the unchanged macOS local Electron receipts remain current, scoped to their exact commands; reduced unit selectors and SSH have no current execution receipt. Historical pre-audit oracle: Synchronously enqueue separate 10,000-entry OSC and DA1 reply floods before the scheduled drain and assert that only the initial immediate reply and newest 64 pending replies are written, each as one provider write, before a trailing keystroke. Stall an acknowledged write between earlier and later ordinary/reply input, then require teardown to settle it and same-id reuse to receive no stale tail. Reenter the queue synchronously from an acknowledged write with both enqueue and clear/reuse, requiring one drain and fresh input delivery only after the stale acknowledged write settles false. At the host boundary, defer an OSC reply and assert that separate or legacy-coalesced DA1/CPR replies flush after it in observed query order. At the remote-runtime boundary, preserve separate writes around pending ordinary input, async validation, and viewport-claim buffering. Repeat behind 10,000 ordinary inputs and exercise the text ceiling, real xterm generation, provider-write failure, rejected yield, and clear/reuse generation fencing.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/input/pty-input-write-queue.test.ts src/renderer/src/components/terminal-pane/pty-transport/ipc/pty-transport-input-write.test.ts src/shared/terminal/protocol/queries/terminal-query-reply.test.ts src/shared/terminal/pty/ingress/pty-startup-ingress-live-query-reply.test.ts src/shared/terminal/pty/ingress/pty-startup-reply-echo-shapes.test.ts", "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/output/remote-runtime-pty-batching.test.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-query-reply-immediate.test.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-input-coalescing.test.ts", @@ -11969,7 +11974,8 @@ "Audit1878 retires this complete dated receipt because a source in its selector was reduced. It does not validate the reduced selector. Original receipt verbatim:\n{\n \"date\": \"2026-08-30\",\n \"runner\": \"local\",\n \"platform\": \"macos\",\n \"command\": \"pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-input-write-queue.test.ts src/renderer/src/components/terminal-pane/pty-transport-input-write.test.ts src/shared/terminal-query-reply.test.ts src/shared/pty-startup-ingress-live-query-reply.test.ts src/shared/pty-startup-reply-echo-shapes.test.ts\",\n \"currentCommand\": \"pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/input/pty-input-write-queue.test.ts src/renderer/src/components/terminal-pane/pty-transport/ipc/pty-transport-input-write.test.ts src/shared/terminal/protocol/queries/terminal-query-reply.test.ts src/shared/terminal/pty/ingress/pty-startup-ingress-live-query-reply.test.ts src/shared/terminal/pty/ingress/pty-startup-reply-echo-shapes.test.ts\",\n \"result\": \"passed\",\n \"durationSeconds\": 15,\n \"summary\": \"Five files and 149 tests passed in 15.16 seconds, including explicit IPC reply-source routing, the 10,000-reply count and text ceilings, 10,000-entry ordinary backlog preservation, acknowledged-write FIFO barriers, synchronous reentrancy fencing, prompt teardown settlement, same-id generation fencing, real xterm OSC query floods, single-shot drain-failure recovery, and one-reply-per-write echo containment.\"\n }", "Audit1878 removes standalone reply-classifier inventory, ordinary-keystroke helper/no-write checks and a latency-order title that never executes the host path from pty-startup-ingress-live-query-reply.test.ts. It also removes projection needle/refusal/holdPartial and ConPTY/WSL string/array inventories from pty-startup-reply-echo-shapes.test.ts, including the direct short-needle and ESC-hold guard alarms. Captured echo matching and actual ingress writes/output state remain, without equivalent replacement coverage. Only the two unchanged macOS local Electron receipts remain active; neither proves the reduced five-file unit selector, SSH, a new soak result, or a new measured budget.", "Audit3086: retire the immediate-reply ceiling and thrown renderer-write recovery aliases. Narrow FIFO coverage to deferred preconnect validation and teardown coverage to natural-exit stale ordinary/accepted tails; live acknowledged barriers and disconnect/detach branches are intentionally retired. Prior queue cleanup and unchanged Electron receipts remain exact; no replacement or fresh execution is claimed.", - "Audit 3292 retires the empty-queue immediate query-reply delivery example from remote-runtime-pty-query-reply-immediate.test.ts. Pending ordinary-input separation and asynchronous paste-validation ordering cases remain byte-identical, matching the existing assertion reference. The no-pending-input delivery alarm is intentionally lost without equivalent replacement. Existing selectors, unchanged local Electron receipts and historical receipts remain exact; no new runtime, PTY deadline, platform or mixed-version result is claimed." + "Audit 3292 retires the empty-queue immediate query-reply delivery example from remote-runtime-pty-query-reply-immediate.test.ts. Pending ordinary-input separation and asynchronous paste-validation ordering cases remain byte-identical, matching the existing assertion reference. The no-pending-input delivery alarm is intentionally lost without equivalent replacement. Existing selectors, unchanged local Electron receipts and historical receipts remain exact; no new runtime, PTY deadline, platform or mixed-version result is claimed.", + "Audit 3312 reduces remote-runtime-pty-batching.test.ts, intentionally losing idle leading/trailing coalescing, leading-edge-disabled, accumulated byte-ceiling flush, standalone synchronous oversized-input rejection, accepted-input validation yield, deferred pending-tail byte order, and queued-viewport clear/flush and timer-cancellation alarms. The unchanged UTF-8 splitting, validation barrier ordering and asynchronous oversized-input nondelivery cases are narrower, not equivalent replacements. The barrier assertion and both existing Electron receipts remain unchanged and do not validate these retired cases. No fresh runtime, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed." ], "demotionRule": "Keep experimental or revert if ordinary input is evicted, pending replies exceed either ceiling, retained replies are coalesced, a trailing key sits behind more than the bounded reply window, or the gate develops unexplained flakes." }, @@ -12736,14 +12742,14 @@ "providers": ["local", "remote-runtime", "ssh"], "coveredPlatforms": ["macos", "linux"], "coveredProviders": ["local", "remote-runtime", "ssh"], - "coverageNotes": "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2056 (sources2054/2055): Intentionally retire these low-value alarms without replacement: src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts: delivers remote file changes to listeners for the same worktree. src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/mobile-session-tabs-part-14.spec'). src/main/runtime/orca-runtime-tests/terminal-side-effect-facts-part-03.spec.ts: arms the stale-title timer for a seeded working title; arms the Command Code scrape from the noted spawn command; prefers the tracked title over a stale renderer lastTitle in the hydration seed; writes recent output into the current runtime output map; uses a later title handler replacement with the runtime receiver; calls the current injected listener with the runtime receiver; keeps optional reconciliation lazy after its callback is removed. src/main/runtime/orca-runtime-tests/mobile-session-tabs-part-14.spec.ts: lists them, so a paired client can mirror a migrated editor; closes one by retiring it from the host session; keeps an unsaved draft unless the close is forced; shares aggregate work but leaves a replacement pending promise alone; lets a targeted refresh finish while an aggregate refresh waits; refuses a close when the store lacks %s before the guard; keeps a later missing %s on the original caught failure path. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Exact retired assertionRefs, preserved as historical objects only: {\n \"file\": \"src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts\",\n \"assertions\": [\n \"terminal watcher error and setup failure evict the shared entry before retrying callbacks run\",\n \"an old stream end cannot evict the replacement stream, which still delivers to its listener\",\n \"two same-worktree listeners share a synthetic subscription until the final stop unsubscribes and sends unwatch\"\n ]\n } Historical prior account (preserved verbatim): Audit1973: Delete the standalone mocked watcher cleanup suite and its entire sole-file prospective command. Retire the removed file-command transport-abort alias; preserve the independent cancellation suite and final-overflow/error/end cleanup callback, and13 unrelated receipt objects. Whole original changed dated receipts are retired, never narrowed or transferred. This metadata review claims no fresh execution, equivalent replacement, provider/PTY, live remote/mobile, rendered, cross-platform, soak or measured-budget result. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Partial runtime files-watch suite retains controlled Windows close-event waiting, shared-client survival, late physical-exit ownership/deadline cleanup and shutdown joining. Its combined Windows close/deadline/late-exit alias remains backed by unchanged callbacks. Retire native-error-close, pre-publication setup ownership and failed-delete rearm aliases; associated delegation, failure forwarding, removal routing and recovery examples are lost. These tests inject watch/subscription seams and do not observe an actual native watcher or physical process. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Partial runtime watcher-host suite loses mocked option/event/overflow/resubscription forwarding, supplied teardown-error propagation, aborted-signal/no-call and private ten-thousand-waiter-count inventories. Retire the 10,000-caller bound and failed-recovery-subscriber-ending assertion aliases. Keep same-root owner cancellation, pre-ready replacement, held root-close join, stale-generation recovery delivery and initial failed-root ownership mappings backed by retained real-host callbacks with injected subscription/error/physical-exit promises. No actual native watcher, process termination or resource bound was observed. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1683 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1645 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1627 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1623 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1516 removes package-electron-runtime-contract.test.mjs, a copied release-workflow platform and marker matrix. Its workflow wiring alarm is intentionally retired; no actual packaged watcher delivery was exercised and no equivalent replacement is claimed. Unaffected dated execution receipts retain their original scope; changed optional selectors have no fresh full-gate execution result. This audit removes the paired-web stopped-during-outage and late-ready direct-branch examples. The retained pre-ready cancellation, failed-unwatch retry and replay-continuity cases are not equivalent replacements for those ordering alarms. Current web-runtime-client.test.ts assertions below name only its retained pre-ready callback case; replay cases remain in their separate suite. Historical commands and results remain historical. Deterministic tests cover one shared healthy child per runtime process plus at most four bounded fault-quarantine children, native desktop, paired-runtime, and WSL snapshot processes sharing the same eight-physical-child reservation with typed serialized event-driven capacity recovery, including recursive re-wait when crash recovery reclaims an announced slot, one quarantine attempt per watch lifetime, generation-scoped cancellation, bounded termination with removable deadline waiters, replacement-crawl errors that cannot be blessed by a late readiness ack, pre-ready paired-web cancellation registered before an unbounded capacity wait, already-resolved and late physical-exit retry of a rejected paired-web teardown, physical desktop-install and local/SSH PTY teardown, stale-generation rejection, per-child stat bounds, final-overflow RPC delivery, and renderer eviction before terminal callbacks can retry. Destructive local and SSH removal fences both ID-derived and resolved-cwd terminal roots, including sibling-root/cwd combinations with reverse admission rollback, closes descendant relay watches before parent deletion, enumerates authoritative provider/cwd ownership, falls back to daemon spawn cwd before OSC 7, and keeps daemon/relay immediate shutdown pending until native exit or a bounded fail-closed error; a dead relay PID discovered during attach also settles concurrent shutdown before its stale entry is reaped, while Windows runtime watcher deadlines retain one removable physical-close waiter, treat Node's error-close path as positive physical-exit proof, and clear root ownership on late close. The relay policy test maps a standard repository's base, Git common directory, and worktree roots to one healthy supervisor, then proves a shared-child failure recovers those roots in separate quarantine supervisors with overflow and resumed events. Abortable capacity, quarantine, runtime-root, relay, SSH, and relay pre-install setup waits attach one reaction to each shared promise and explicitly remove 10,000 cancelled caller closures while one anchor remains. Built-entry macOS harnesses kill the desktop/runtime and SSH relay watcher children, then require automatic resubscription and later events while the host process and relay PTY survive. SSH bundles the child boundary beside relay.js, requires both artifacts for install completeness, and preserves registration-owned same-root cancellation. The test-value audit corrects the active renderer watch command/file reference to the existing runtime-file-client-watch.test.ts, whose three cases are retained unchanged. They observe synthetic subscription retry/fanout and unsubscribe/unwatch calls, not actual preload/main bookkeeping or physical watcher-process cleanup. The separate read/preview client suite loses forwarding, supplied-error and negative-only examples, retaining binary detection and Unicode SVG decoding. Historical evidence commands/results are unchanged. Test-value audit1293 removes src/main/providers/local-pty/local-pty-provider-session-inventory.test.ts (2 authored cases). Their isolated assertions no longer run; surviving tests are not claimed as equivalent coverage. Fresh execution against the candidate SHA is required. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1412: src/main/workspaces/removal/watcher-removal-gate.test.ts reduced. Loses exact removal-error compatibility with renderer predicate under shared message constants. Kept public refusal outcomes do not establish that renderer classification. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result.", + "coverageNotes": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2056 (sources2054/2055): Intentionally retire these low-value alarms without replacement: src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts: delivers remote file changes to listeners for the same worktree. src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/mobile-session-tabs-part-14.spec'). src/main/runtime/orca-runtime-tests/terminal-side-effect-facts-part-03.spec.ts: arms the stale-title timer for a seeded working title; arms the Command Code scrape from the noted spawn command; prefers the tracked title over a stale renderer lastTitle in the hydration seed; writes recent output into the current runtime output map; uses a later title handler replacement with the runtime receiver; calls the current injected listener with the runtime receiver; keeps optional reconciliation lazy after its callback is removed. src/main/runtime/orca-runtime-tests/mobile-session-tabs-part-14.spec.ts: lists them, so a paired client can mirror a migrated editor; closes one by retiring it from the host session; keeps an unsaved draft unless the close is forced; shares aggregate work but leaves a replacement pending promise alone; lets a targeted refresh finish while an aggregate refresh waits; refuses a close when the store lacks %s before the guard; keeps a later missing %s on the original caught failure path. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Exact retired assertionRefs, preserved as historical objects only: {\n \"file\": \"src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts\",\n \"assertions\": [\n \"terminal watcher error and setup failure evict the shared entry before retrying callbacks run\",\n \"an old stream end cannot evict the replacement stream, which still delivers to its listener\",\n \"two same-worktree listeners share a synthetic subscription until the final stop unsubscribes and sends unwatch\"\n ]\n } Historical prior account (preserved verbatim): Audit1973: Delete the standalone mocked watcher cleanup suite and its entire sole-file prospective command. Retire the removed file-command transport-abort alias; preserve the independent cancellation suite and final-overflow/error/end cleanup callback, and13 unrelated receipt objects. Whole original changed dated receipts are retired, never narrowed or transferred. This metadata review claims no fresh execution, equivalent replacement, provider/PTY, live remote/mobile, rendered, cross-platform, soak or measured-budget result. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Partial runtime files-watch suite retains controlled Windows close-event waiting, shared-client survival, late physical-exit ownership/deadline cleanup and shutdown joining. Its combined Windows close/deadline/late-exit alias remains backed by unchanged callbacks. Retire native-error-close, pre-publication setup ownership and failed-delete rearm aliases; associated delegation, failure forwarding, removal routing and recovery examples are lost. These tests inject watch/subscription seams and do not observe an actual native watcher or physical process. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Partial runtime watcher-host suite loses mocked option/event/overflow/resubscription forwarding, supplied teardown-error propagation, aborted-signal/no-call and private ten-thousand-waiter-count inventories. Retire the 10,000-caller bound and failed-recovery-subscriber-ending assertion aliases. Keep same-root owner cancellation, pre-ready replacement, held root-close join, stale-generation recovery delivery and initial failed-root ownership mappings backed by retained real-host callbacks with injected subscription/error/physical-exit promises. No actual native watcher, process termination or resource bound was observed. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1683 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1645 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1627 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1623 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1516 removes package-electron-runtime-contract.test.mjs, a copied release-workflow platform and marker matrix. Its workflow wiring alarm is intentionally retired; no actual packaged watcher delivery was exercised and no equivalent replacement is claimed. Unaffected dated execution receipts retain their original scope; changed optional selectors have no fresh full-gate execution result. This audit removes the paired-web stopped-during-outage and late-ready direct-branch examples. The retained pre-ready cancellation, failed-unwatch retry and replay-continuity cases are not equivalent replacements for those ordering alarms. Current web-runtime-client.test.ts assertions below name only its retained pre-ready callback case; replay cases remain in their separate suite. Historical commands and results remain historical. Deterministic tests cover one shared healthy child per runtime process plus at most four bounded fault-quarantine children, native desktop, paired-runtime, and WSL snapshot processes sharing the same eight-physical-child reservation with typed serialized event-driven capacity recovery, including recursive re-wait when crash recovery reclaims an announced slot, one quarantine attempt per watch lifetime, generation-scoped cancellation, bounded termination with removable deadline waiters, replacement-crawl errors that cannot be blessed by a late readiness ack, pre-ready paired-web cancellation registered before an unbounded capacity wait, already-resolved and late physical-exit retry of a rejected paired-web teardown, physical desktop-install and local/SSH PTY teardown, stale-generation rejection, per-child stat bounds, final-overflow RPC delivery, and renderer eviction before terminal callbacks can retry. Destructive local and SSH removal fences both ID-derived and resolved-cwd terminal roots, including sibling-root/cwd combinations with reverse admission rollback, closes descendant relay watches before parent deletion, enumerates authoritative provider/cwd ownership, falls back to daemon spawn cwd before OSC 7, and keeps daemon/relay immediate shutdown pending until native exit or a bounded fail-closed error; a dead relay PID discovered during attach also settles concurrent shutdown before its stale entry is reaped, while Windows runtime watcher deadlines retain one removable physical-close waiter, treat Node's error-close path as positive physical-exit proof, and clear root ownership on late close. The relay policy test maps a standard repository's base, Git common directory, and worktree roots to one healthy supervisor, then proves a shared-child failure recovers those roots in separate quarantine supervisors with overflow and resumed events. Abortable capacity, quarantine, runtime-root, relay, SSH, and relay pre-install setup waits attach one reaction to each shared promise and explicitly remove 10,000 cancelled caller closures while one anchor remains. Built-entry macOS harnesses kill the desktop/runtime and SSH relay watcher children, then require automatic resubscription and later events while the host process and relay PTY survive. SSH bundles the child boundary beside relay.js, requires both artifacts for install completeness, and preserves registration-owned same-root cancellation. The test-value audit corrects the active renderer watch command/file reference to the existing runtime-file-client-watch.test.ts, whose three cases are retained unchanged. They observe synthetic subscription retry/fanout and unsubscribe/unwatch calls, not actual preload/main bookkeeping or physical watcher-process cleanup. The separate read/preview client suite loses forwarding, supplied-error and negative-only examples, retaining binary detection and Unicode SVG decoding. Historical evidence commands/results are unchanged. Test-value audit1293 removes src/main/providers/local-pty/local-pty-provider-session-inventory.test.ts (2 authored cases). Their isolated assertions no longer run; surviving tests are not claimed as equivalent coverage. Fresh execution against the candidate SHA is required. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1412: src/main/workspaces/removal/watcher-removal-gate.test.ts reduced. Loses exact removal-error compatibility with renderer predicate under shared message constants. Kept public refusal outcomes do not establish that renderer classification. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result.", "motivatingLinks": [ "https://github.com/stablyai/orca/issues/5308", "https://github.com/stablyai/orca/issues/7547", "https://github.com/stablyai/orca/issues/8212" ], "invariant": "Local runtime and SSH relay roots share one watcher child per host process while healthy and use at most four quarantine children after shard failure; native desktop, paired-runtime, and direct WSL snapshot children consume the same process-wide eight-child reservation and retry transient exhaustion without permanently disabling a root. Cancellation remains owned by its child generation and settles only after acknowledgement or bounded physical exit; paired-web retains cleanup ownership and can cancel before readiness even while capacity is exhausted, retries a failed unwatch after already-resolved or late physical exit without duplicating stream termination, replacements never overlap a still-retiring generation, and aborted or timed-out callers detach immediately even while queued behind same-root relay teardown. Destructive local and SSH cleanup holds one containment-aware watcher/terminal removal fence through descendant watcher and PTY physical exit plus parent Git/filesystem deletion; daemon and relay kill requests retain ownership until the actual subprocess exit callback or an equivalent positive dead-PID proof and propagate rejected signals or a bounded failure otherwise. A crash-fused shared shard moves affected roots into the bounded quarantine pool, a second fused quarantine ends those watch lifetimes, interrupted setup emits overflow after recovery, metadata work stays bounded, and the SSH relay, PTYs, Git, and other filesystem operations never load the watcher native addon or share its fault boundary.", - "oracle": "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2056 (sources2054/2055): Intentionally retire these low-value alarms without replacement: src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts: delivers remote file changes to listeners for the same worktree. src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/mobile-session-tabs-part-14.spec'). src/main/runtime/orca-runtime-tests/terminal-side-effect-facts-part-03.spec.ts: arms the stale-title timer for a seeded working title; arms the Command Code scrape from the noted spawn command; prefers the tracked title over a stale renderer lastTitle in the hydration seed; writes recent output into the current runtime output map; uses a later title handler replacement with the runtime receiver; calls the current injected listener with the runtime receiver; keeps optional reconciliation lazy after its callback is removed. src/main/runtime/orca-runtime-tests/mobile-session-tabs-part-14.spec.ts: lists them, so a paired client can mirror a migrated editor; closes one by retiring it from the host session; keeps an unsaved draft unless the close is forced; shares aggregate work but leaves a replacement pending promise alone; lets a targeted refresh finish while an aggregate refresh waits; refuses a close when the store lacks %s before the guard; keeps a later missing %s on the original caught failure path. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Exact retired assertionRefs, preserved as historical objects only: {\n \"file\": \"src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts\",\n \"assertions\": [\n \"terminal watcher error and setup failure evict the shared entry before retrying callbacks run\",\n \"an old stream end cannot evict the replacement stream, which still delivers to its listener\",\n \"two same-worktree listeners share a synthetic subscription until the final stop unsubscribes and sends unwatch\"\n ]\n } Historical prior account (preserved verbatim): Audit1973: Delete the standalone mocked watcher cleanup suite and its entire sole-file prospective command. Retire the removed file-command transport-abort alias; preserve the independent cancellation suite and final-overflow/error/end cleanup callback, and13 unrelated receipt objects. Whole original changed dated receipts are retired, never narrowed or transferred. This metadata review claims no fresh execution, equivalent replacement, provider/PTY, live remote/mobile, rendered, cross-platform, soak or measured-budget result. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Partial runtime files-watch suite retains controlled Windows close-event waiting, shared-client survival, late physical-exit ownership/deadline cleanup and shutdown joining. Its combined Windows close/deadline/late-exit alias remains backed by unchanged callbacks. Retire native-error-close, pre-publication setup ownership and failed-delete rearm aliases; associated delegation, failure forwarding, removal routing and recovery examples are lost. These tests inject watch/subscription seams and do not observe an actual native watcher or physical process. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Partial runtime watcher-host suite loses mocked option/event/overflow/resubscription forwarding, supplied teardown-error propagation, aborted-signal/no-call and private ten-thousand-waiter-count inventories. Retire the 10,000-caller bound and failed-recovery-subscriber-ending assertion aliases. Keep same-root owner cancellation, pre-ready replacement, held root-close join, stale-generation recovery delivery and initial failed-root ownership mappings backed by retained real-host callbacks with injected subscription/error/physical-exit promises. No actual native watcher, process termination or resource bound was observed. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1683 retires changed-source assertion and execution scope without equivalent replacement coverage. Existing unchanged-selector receipts remain scoped to their original sources. Historical prior-source oracle: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Fill the global watcher-child reservation, require native desktop, paired-runtime, and WSL setup to retain typed transient capacity and retry after one serialized release, then cancel pending desktop and paired-client watches and require no retry. Hold one live anchor while aborting 10,000 callers at the capacity, quarantine-assignment, runtime-root, relay-setup, and SSH-registration layers; require exactly one explicit waiter and one base-promise reaction to remain, then zero after settlement. Hold an old relay unsubscribe, queue one same-root setup leader and one anchor, then abort 10,000 joiners; require every cancellation to settle before teardown releases, exactly one anchor waiter, and only one later physical install. Advance the former paired-web callback deadline while capacity remains full, require cancellation ownership to stay live, publish the server token before ready, and require files.unwatch to abort setup before a later slot release can install a native root. Reject the first ready-watch unwatch with a physical-exit promise, resolve that exit later, and require a fresh physical unwatch attempt, registry release, and exactly one stream end. Fill one healthy shared child with multiple roots, map a standard SSH repository's base, Git common directory, and worktree roots to one relay supervisor, then inject IPC disconnect-before-exit, queued and active cancellation, operations during termination, kill-without-exit, repeated unkillable generations, initial and recovery timeouts, and shard failure. During local and SSH deletion, attempt a late terminal spawn whose worktree identity and cwd name different roots, enumerate a non-prefix remote PTY by authoritative worktree ownership, withhold daemon/relay exit callbacks, make attach positively prove a concurrent-shutdown PID dead, and make the provider unavailable; require every acquired admission to roll back on rejection, dead-PID proof to settle physical exit before reaping, timeout to retain ownership, and Git/filesystem deletion never to start. Require bounded watcher process count, physical-exit settlement, healthy-root recovery, stale-generation rejection, overflow plus later events, and fail-closed destructive cleanup. Verify relay build hashes and remote install probes require both relay.js and relay-watcher.js. Launch the built relay with a real PTY and Parcel watch, kill only the watcher child, and require the same relay PID and PTY to answer before a replacement child delivers overflow and a post-crash event.", + "oracle": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2056 (sources2054/2055): Intentionally retire these low-value alarms without replacement: src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts: delivers remote file changes to listeners for the same worktree. src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/mobile-session-tabs-part-14.spec'). src/main/runtime/orca-runtime-tests/terminal-side-effect-facts-part-03.spec.ts: arms the stale-title timer for a seeded working title; arms the Command Code scrape from the noted spawn command; prefers the tracked title over a stale renderer lastTitle in the hydration seed; writes recent output into the current runtime output map; uses a later title handler replacement with the runtime receiver; calls the current injected listener with the runtime receiver; keeps optional reconciliation lazy after its callback is removed. src/main/runtime/orca-runtime-tests/mobile-session-tabs-part-14.spec.ts: lists them, so a paired client can mirror a migrated editor; closes one by retiring it from the host session; keeps an unsaved draft unless the close is forced; shares aggregate work but leaves a replacement pending promise alone; lets a targeted refresh finish while an aggregate refresh waits; refuses a close when the store lacks %s before the guard; keeps a later missing %s on the original caught failure path. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Exact retired assertionRefs, preserved as historical objects only: {\n \"file\": \"src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts\",\n \"assertions\": [\n \"terminal watcher error and setup failure evict the shared entry before retrying callbacks run\",\n \"an old stream end cannot evict the replacement stream, which still delivers to its listener\",\n \"two same-worktree listeners share a synthetic subscription until the final stop unsubscribes and sends unwatch\"\n ]\n } Historical prior account (preserved verbatim): Audit1973: Delete the standalone mocked watcher cleanup suite and its entire sole-file prospective command. Retire the removed file-command transport-abort alias; preserve the independent cancellation suite and final-overflow/error/end cleanup callback, and13 unrelated receipt objects. Whole original changed dated receipts are retired, never narrowed or transferred. This metadata review claims no fresh execution, equivalent replacement, provider/PTY, live remote/mobile, rendered, cross-platform, soak or measured-budget result. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Partial runtime files-watch suite retains controlled Windows close-event waiting, shared-client survival, late physical-exit ownership/deadline cleanup and shutdown joining. Its combined Windows close/deadline/late-exit alias remains backed by unchanged callbacks. Retire native-error-close, pre-publication setup ownership and failed-delete rearm aliases; associated delegation, failure forwarding, removal routing and recovery examples are lost. These tests inject watch/subscription seams and do not observe an actual native watcher or physical process. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Partial runtime watcher-host suite loses mocked option/event/overflow/resubscription forwarding, supplied teardown-error propagation, aborted-signal/no-call and private ten-thousand-waiter-count inventories. Retire the 10,000-caller bound and failed-recovery-subscriber-ending assertion aliases. Keep same-root owner cancellation, pre-ready replacement, held root-close join, stale-generation recovery delivery and initial failed-root ownership mappings backed by retained real-host callbacks with injected subscription/error/physical-exit promises. No actual native watcher, process termination or resource bound was observed. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1683 retires changed-source assertion and execution scope without equivalent replacement coverage. Existing unchanged-selector receipts remain scoped to their original sources. Historical prior-source oracle: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Fill the global watcher-child reservation, require native desktop, paired-runtime, and WSL setup to retain typed transient capacity and retry after one serialized release, then cancel pending desktop and paired-client watches and require no retry. Hold one live anchor while aborting 10,000 callers at the capacity, quarantine-assignment, runtime-root, relay-setup, and SSH-registration layers; require exactly one explicit waiter and one base-promise reaction to remain, then zero after settlement. Hold an old relay unsubscribe, queue one same-root setup leader and one anchor, then abort 10,000 joiners; require every cancellation to settle before teardown releases, exactly one anchor waiter, and only one later physical install. Advance the former paired-web callback deadline while capacity remains full, require cancellation ownership to stay live, publish the server token before ready, and require files.unwatch to abort setup before a later slot release can install a native root. Reject the first ready-watch unwatch with a physical-exit promise, resolve that exit later, and require a fresh physical unwatch attempt, registry release, and exactly one stream end. Fill one healthy shared child with multiple roots, map a standard SSH repository's base, Git common directory, and worktree roots to one relay supervisor, then inject IPC disconnect-before-exit, queued and active cancellation, operations during termination, kill-without-exit, repeated unkillable generations, initial and recovery timeouts, and shard failure. During local and SSH deletion, attempt a late terminal spawn whose worktree identity and cwd name different roots, enumerate a non-prefix remote PTY by authoritative worktree ownership, withhold daemon/relay exit callbacks, make attach positively prove a concurrent-shutdown PID dead, and make the provider unavailable; require every acquired admission to roll back on rejection, dead-PID proof to settle physical exit before reaping, timeout to retain ownership, and Git/filesystem deletion never to start. Require bounded watcher process count, physical-exit settlement, healthy-root recovery, stale-generation rejection, overflow plus later events, and fail-closed destructive cleanup. Verify relay build hashes and remote install probes require both relay.js and relay-watcher.js. Launch the built relay with a real PTY and Parcel watch, kill only the watcher child, and require the same relay PID and PTY to answer before a replacement child delivers overflow and a post-crash event.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-unsubscribe-timeout.test.ts --reporter=dot", "pnpm exec vitest run --config config/vitest.config.ts src/main/ipc/filesystem/watchers/filesystem-watcher.test.ts src/main/ipc/filesystem/watchers/filesystem-watcher-native-capacity.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-child-registry.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-child-launch.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-process.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-supervisor-capacity.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-disconnect-termination.test.ts src/main/ipc/filesystem/watcher-process/runtime-watcher-pending-assignment.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-event-delivery.test.ts src/main/ipc/filesystem/watchers/filesystem-watcher-local-unsubscribe.test.ts src/main/ipc/filesystem/watchers/filesystem-watcher-wsl.test.ts src/main/ipc/worktrees-delete-pty-teardown.test.ts src/main/ipc/worktrees-orphan-directory-cleanup.test.ts src/main/ipc/worktrees-remove-preflight.test.ts src/main/workspaces/removal/watcher-removal-gate.test.ts src/main/runtime/files/watch/orca-runtime-files-watch.test.ts src/main/runtime/workspaces/removal/worktree-teardown.test.ts src/main/runtime/construction/orca-runtime.test.ts src/main/runtime/rpc/methods/files/files.test.ts src/main/runtime/rpc/methods/files/files-watch-cancellation.test.ts src/renderer/src/web/web-runtime-client.test.ts src/shared/cross-platform-path.test.ts src/shared/physical-exit-tracker.test.ts src/shared/promise-settlement-waiters.test.ts src/main/ipc/pty/pty-spawn-cwd-fallback.test.ts src/main/providers/local-pty/local-pty-provider-shutdown.test.ts src/main/daemon/providers/adapter/testing/daemon-pty-adapter.test.ts src/main/daemon/pty-subprocess/testing/pty-subprocess.test.ts src/main/daemon/sessions/testing/session.test.ts src/main/daemon/terminal-host/testing/terminal-host.test.ts src/main/daemon/terminal-host/testing/terminal-host-session-reaping-leak.test.ts --reporter=dot", @@ -13182,7 +13188,8 @@ "Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims.", "Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical.", "Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed.", - "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed." + "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed.", + "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed." ], "demotionRule": "Demote or quarantine if the fault harness flakes without a product or harness bug, if healthy operation exceeds one runtime watcher child, if total physical operation exceeds eight children including retiring generations, if quarantine children outlive their roots or repeat after fusing, if event delivery becomes unbounded, or if metadata/stat work returns to the serve process." }, @@ -14427,10 +14434,10 @@ "providers": ["paired-runtime"], "coveredPlatforms": ["linux"], "coveredProviders": ["paired-runtime"], - "coverageNotes": "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit1579 removes workflow/manifest/regex/private binary-header and diagnostic metric inventories, fixture selectors, injected forwarding/platform/signature/process examples, unproven overlap and internal churn counters without equivalent replacement coverage. Retains actual Git trust/diffs; filesystem hash/ELF-region/native discovery; child CLI module repair; compiler/lock/publication/rollback and synthetic Electron install/cache ownership; executable shell quoting/exit; controlled notary/timer cleanup; local patched extractor file/AST freshness; real Chromium/WebKit CSP/sandbox/HTTPS image/referrer/link/preservation controls under parsed Android headers. Five unchanged setup-pnpm/VM request-cleanup definitions reviewed and not rerun. GitHub, installer/archive, OS metric, client, native delegate, compressor, builder and Apple boundaries injected or absent where recorded. No actual native PTY, shipped full/native app, real Actions/cache/npm signature authority, downloaded Electron archive, real objdump/linker/native CLI/AppImage, signing/NSIS/notary/upload/updater or live freeze proof. Historical scope before this audit: Published Linux 1.4.188 desktop against current source in both directions; scheduled weekly and manually runnable. No required PR check.", + "coverageNotes": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit1579 removes workflow/manifest/regex/private binary-header and diagnostic metric inventories, fixture selectors, injected forwarding/platform/signature/process examples, unproven overlap and internal churn counters without equivalent replacement coverage. Retains actual Git trust/diffs; filesystem hash/ELF-region/native discovery; child CLI module repair; compiler/lock/publication/rollback and synthetic Electron install/cache ownership; executable shell quoting/exit; controlled notary/timer cleanup; local patched extractor file/AST freshness; real Chromium/WebKit CSP/sandbox/HTTPS image/referrer/link/preservation controls under parsed Android headers. Five unchanged setup-pnpm/VM request-cleanup definitions reviewed and not rerun. GitHub, installer/archive, OS metric, client, native delegate, compressor, builder and Apple boundaries injected or absent where recorded. No actual native PTY, shipped full/native app, real Actions/cache/npm signature authority, downloaded Electron archive, real objdump/linker/native CLI/AppImage, signing/NSIS/notary/upload/updater or live freeze proof. Historical scope before this audit: Published Linux 1.4.188 desktop against current source in both directions; scheduled weekly and manually runnable. No required PR check.", "motivatingLinks": ["https://github.com/stablyai/orca/actions/runs/34069063016"], "invariant": "A paired client and host without client-hosted browser capabilities retain server-hosted browser placement across supported version skew.", - "oracle": "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Require both existing named browser placement scenarios to pass three times with one attempt, zero skips, zero failures, and no report errors.", + "oracle": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Require both existing named browser placement scenarios to pass three times with one attempt, zero skips, zero failures, and no report errors.", "commands": [ "gh workflow run packaged-browser-e2e.yml", "pnpm exec playwright test tests/e2e/packaged-mixed-version-browser-placement.spec.ts --config tests/playwright.config.ts --project=electron-headless --workers=1 --repeat-each=3 --retries=0", @@ -14517,7 +14524,8 @@ "Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims.", "Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical.", "Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed.", - "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed." + "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed.", + "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed." ], "demotionRule": "Keep experimental if any direction skips or fails; do not extend timeouts or retry to green." }, diff --git a/config/scripts/ci/sharding/ci-shard-timings.json b/config/scripts/ci/sharding/ci-shard-timings.json index 0c336f13bd1..c468fdc0660 100644 --- a/config/scripts/ci/sharding/ci-shard-timings.json +++ b/config/scripts/ci/sharding/ci-shard-timings.json @@ -6124,7 +6124,6 @@ "src/renderer/src/hooks/useIpcEvents-browser-certificate-failure.test.ts": 2981, "src/renderer/src/hooks/useIpcEvents-browser-navigation.test.ts": 1511, "src/renderer/src/hooks/useIpcEvents-client-hosted-browser-rows.test.ts": 1894, - "src/renderer/src/hooks/useIpcEvents-close-routing-active-browser-tab.test.ts": 4572, "src/renderer/src/hooks/useIpcEvents-close-routing-session-tabs.test.ts": 5812, "src/renderer/src/hooks/useIpcEvents-lifecycle.test.ts": 2326, "src/renderer/src/hooks/useIpcEvents-new-workspace-shortcut.test.ts": 72, @@ -6404,7 +6403,6 @@ "src/renderer/src/lib/workspace/activation/worktree-activation-structured-chat-surface.test.ts": 2043, "src/renderer/src/lib/workspace/activation/worktree-activation-web-runtime.test.ts": 2768, "src/renderer/src/lib/workspace/activation/worktree-agent-activation-seam.test.ts": 3440, - "src/renderer/src/lib/workspace/creation/worktree-creation-backend-startup-focus.test.ts": 2267, "src/renderer/src/lib/workspace/creation/worktree-creation-initial-surface.test.ts": 5258, "src/renderer/src/lib/workspace/creation/worktree-creation-structured-session-lifecycle.test.ts": 2633, "src/renderer/src/lib/workspace/creation/worktree-creation-structured-session-prompts.test.tsx": 4554, @@ -6578,7 +6576,6 @@ "src/renderer/src/store/slices/browser/cleanup/browser-cleanup-close.test.ts": 1911, "src/renderer/src/store/slices/browser/cleanup/browser-mount-admission-removal.test.ts": 1982, "src/renderer/src/store/slices/browser/pages/browser-page-conversion.test.ts": 1777, - "src/renderer/src/store/slices/browser/hosts/browser-remote-tab-creation.test.ts": 2119, "src/renderer/src/store/slices/browser/hosts/browser-session-host-selection.test.ts": 988, "src/renderer/src/store/slices/browser/profiles/browser-session-profiles.test.ts": 792, "src/renderer/src/store/slices/browser/tabs/browser-tab-creation-placement.test.ts": 1886, diff --git a/src/renderer/src/components/terminal-pane/provider-notices/agent-presence/agent-terminal-presence.test.ts b/src/renderer/src/components/terminal-pane/provider-notices/agent-presence/agent-terminal-presence.test.ts deleted file mode 100644 index 62b44a4afba..00000000000 --- a/src/renderer/src/components/terminal-pane/provider-notices/agent-presence/agent-terminal-presence.test.ts +++ /dev/null @@ -1,40 +0,0 @@ -import { afterEach, describe, expect, it, vi } from 'vitest' -import { useAppStore } from '@/store/index' -import { makeTerminalTab } from '@/store/terminals/terminal-row.test-fixture' -import { status } from '@/components/native-chat/messages/agent-senders/native-chat-agent-sender.test-fixture' -import { whenAgentTerminalAppears } from './agent-terminal-presence' - -const initial = useAppStore.getState() - -afterEach(() => { - useAppStore.setState(initial, true) -}) - -describe('whenAgentTerminalAppears', () => { - it('ignores a Claude status entry from a native chat tab', () => { - const onAppear = vi.fn() - useAppStore.setState({ - tabsByWorktree: {}, - agentStatusByPaneKey: {}, - paneForegroundAgentByPaneKey: {} - }) - const unsubscribe = whenAgentTerminalAppears('claude', onAppear) - const entry = status(1) - useAppStore.setState({ agentStatusByPaneKey: { [entry.paneKey]: entry } }) - expect(onAppear).not.toHaveBeenCalled() - unsubscribe() - }) - - it('fires for a Claude status entry from a terminal tab', () => { - const onAppear = vi.fn() - useAppStore.setState({ - tabsByWorktree: { workspace: [makeTerminalTab({ id: 'tab-1', worktreeId: 'workspace' })] }, - agentStatusByPaneKey: {}, - paneForegroundAgentByPaneKey: {} - }) - whenAgentTerminalAppears('claude', onAppear) - const entry = status(1) - useAppStore.setState({ agentStatusByPaneKey: { [entry.paneKey]: entry } }) - expect(onAppear).toHaveBeenCalledTimes(1) - }) -}) diff --git a/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/output/remote-runtime-pty-batching.test.ts b/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/output/remote-runtime-pty-batching.test.ts index fed1cfe15b5..875af4f6904 100644 --- a/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/output/remote-runtime-pty-batching.test.ts +++ b/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/output/remote-runtime-pty-batching.test.ts @@ -1,83 +1,9 @@ import { describe, expect, it, vi } from 'vitest' import { CLIPBOARD_TEXT_MEASURE_YIELD_CODE_UNITS } from '../../../../../../../shared/clipboard-text' -import { - createRemoteRuntimePtyTextBatcher, - createRemoteRuntimeViewportBatcher -} from './remote-runtime-pty-batching' +import { createRemoteRuntimePtyTextBatcher } from './remote-runtime-pty-batching' describe('createRemoteRuntimePtyTextBatcher', () => { - it('sends idle input at once and coalesces input inside the window', async () => { - vi.useFakeTimers() - try { - const flushes: string[] = [] - const batcher = createRemoteRuntimePtyTextBatcher(10, (text) => flushes.push(text), { - maxPendingBytes: 8 - }) - - expect(batcher.push('a')).toBe(true) - expect(flushes).toEqual(['a']) - expect(batcher.push('b')).toBe(true) - expect(batcher.push('c')).toBe(true) - expect(flushes).toEqual(['a']) - // Why: a remount during the window discards these, and the pane must know to say so. - expect(batcher.hasPendingInput()).toBe(true) - - await vi.advanceTimersByTimeAsync(10) - expect(flushes).toEqual(['a', 'bc']) - expect(batcher.hasPendingInput()).toBe(false) - - // The window stays open after a trailing flush; once it lapses idle, the next key goes at once. - expect(batcher.push('d')).toBe(true) - expect(flushes).toEqual(['a', 'bc']) - await vi.advanceTimersByTimeAsync(10) - expect(flushes).toEqual(['a', 'bc', 'd']) - await vi.advanceTimersByTimeAsync(10) - expect(batcher.push('e')).toBe(true) - expect(flushes).toEqual(['a', 'bc', 'd', 'e']) - } finally { - vi.useRealTimers() - } - }) - - it('keeps idle input trailing while the leading edge is disabled', () => { - vi.useFakeTimers() - try { - const flushes: string[] = [] - const batcher = createRemoteRuntimePtyTextBatcher(10, (text) => flushes.push(text), { - sendLeadingEdge: () => false - }) - - expect(batcher.push('a')).toBe(true) - expect(batcher.push('b')).toBe(true) - expect(flushes).toEqual([]) - vi.advanceTimersByTime(10) - expect(flushes).toEqual(['ab']) - } finally { - vi.useRealTimers() - } - }) - - it('flushes before pending input exceeds the byte ceiling', () => { - vi.useFakeTimers() - try { - const flushes: string[] = [] - const batcher = createRemoteRuntimePtyTextBatcher(10, (text) => flushes.push(text), { - maxPendingBytes: 4 - }) - - expect(batcher.push('x')).toBe(true) - expect(batcher.push('ab')).toBe(true) - expect(batcher.push('cd')).toBe(true) - expect(flushes).toEqual(['x']) - - expect(batcher.push('e')).toBe(true) - expect(flushes).toEqual(['x', 'abcd', 'e']) - } finally { - vi.useRealTimers() - } - }) - it('splits one large UTF-8 input without splitting a code point', () => { const flushes: string[] = [] const batcher = createRemoteRuntimePtyTextBatcher(10, (text) => flushes.push(text), { @@ -91,64 +17,6 @@ describe('createRemoteRuntimePtyTextBatcher', () => { expect(flushes).toEqual(['ab', '😀', 'cd']) }) - it('rejects oversized input without flushing clipboard content', () => { - const flushes: string[] = [] - const batcher = createRemoteRuntimePtyTextBatcher(10, (text) => flushes.push(text), { - maxBytes: 4, - maxPendingBytes: 4 - }) - - const secret = ['secret-token-', 'payload'].join('') - - expect(batcher.push(secret)).toBe(false) - expect(flushes).toEqual([]) - }) - - it('yields while validating large accepted input before sending it', async () => { - vi.useFakeTimers() - try { - const flushes: string[] = [] - const text = 'x'.repeat(CLIPBOARD_TEXT_MEASURE_YIELD_CODE_UNITS + 1) - const batcher = createRemoteRuntimePtyTextBatcher(10, (value) => flushes.push(value), { - maxPendingBytes: text.length + 1 - }) - - expect(batcher.push(text)).toBe(true) - expect(flushes).toEqual([]) - - const drained = batcher.drain() - await vi.advanceTimersByTimeAsync(0) - await drained - - expect(flushes).toEqual([text]) - } finally { - vi.useRealTimers() - } - }) - - it('keeps input queued after deferred validation in byte order', async () => { - vi.useFakeTimers() - try { - const flushes: string[] = [] - const text = 'x'.repeat(CLIPBOARD_TEXT_MEASURE_YIELD_CODE_UNITS + 1) - const batcher = createRemoteRuntimePtyTextBatcher(1_000, (value) => flushes.push(value), { - maxPendingBytes: text.length + 10 - }) - - expect(batcher.push(text)).toBe(true) - expect(batcher.push('tail')).toBe(true) - - const drained = batcher.drain() - await vi.advanceTimersByTimeAsync(0) - await drained - - expect(flushes).toEqual([text]) - expect(batcher.takePending()).toBe('tail') - } finally { - vi.useRealTimers() - } - }) - it('runs a validation barrier before input queued after it', async () => { vi.useFakeTimers() try { @@ -202,42 +70,3 @@ describe('createRemoteRuntimePtyTextBatcher', () => { } }) }) - -describe('createRemoteRuntimeViewportBatcher', () => { - it('drops the queued viewport on clear so a later flush emits nothing', () => { - vi.useFakeTimers() - try { - const resizes: { cols: number; rows: number }[] = [] - const batcher = createRemoteRuntimeViewportBatcher(33, (cols, rows) => { - resizes.push({ cols, rows }) - }) - - batcher.queue(120, 40) - batcher.clear() - // A stale pending viewport left behind by clear() would leak out here. - batcher.flush() - - expect(resizes).toEqual([]) - } finally { - vi.useRealTimers() - } - }) - - it('does not emit a cleared viewport when the debounce timer would have fired', () => { - vi.useFakeTimers() - try { - const resizes: { cols: number; rows: number }[] = [] - const batcher = createRemoteRuntimeViewportBatcher(33, (cols, rows) => { - resizes.push({ cols, rows }) - }) - - batcher.queue(90, 30) - batcher.clear() - vi.advanceTimersByTime(100) - - expect(resizes).toEqual([]) - } finally { - vi.useRealTimers() - } - }) -}) diff --git a/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts b/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts index c18b857a7f8..b93d6c6594a 100644 --- a/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts +++ b/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts @@ -139,28 +139,6 @@ describe('paired client sleep/wake with several remote terminal tabs', () => { return panes } - it('rebinds every pane after the shared multiplex connection drops', async () => { - const panes = await attachAllPanes() - const before = subscribedHandles().length - - // Laptop closed: the single multiplex socket dies, taking all six streams. - subscriptionCallbacks?.onClose?.() - - await vi.waitFor(() => - expect(subscribedHandles().length).toBeGreaterThanOrEqual(before + PANE_COUNT) - ) - const afterReconnect = subscribedHandles().slice(before) - expect(new Set(afterReconnect).size, `resubscribes: ${JSON.stringify(afterReconnect)}`).toBe( - PANE_COUNT - ) - for (const pane of panes) { - expect(pane.transport.getPtyId(), `pane ${pane.hostTabId} bound the wrong handle`).toBe( - `remote:env-1@@${hostHandleByTabId.get(pane.hostTabId)}` - ) - pane.transport.destroy?.() - } - }) - it('re-materializes a pane whose host PTY was parked while the client slept', async () => { const panes = await attachAllPanes() const before = subscribedHandles().length diff --git a/src/renderer/src/hooks/useIpcEvents-close-routing-active-browser-tab.test.ts b/src/renderer/src/hooks/useIpcEvents-close-routing-active-browser-tab.test.ts deleted file mode 100644 index c22f0c75ccb..00000000000 --- a/src/renderer/src/hooks/useIpcEvents-close-routing-active-browser-tab.test.ts +++ /dev/null @@ -1,188 +0,0 @@ -import { beforeEach, describe, expect, it, vi } from 'vitest' -import { - useIpcEventsForCloseRouting, - type CloseActiveTabListener -} from './ipc-events-close-routing-test-harness' - -const { closeWebRuntimeSessionTab, destroyWorkspaceWebviews, isWebRuntimeSessionActive } = - vi.hoisted(() => ({ - closeWebRuntimeSessionTab: vi.fn(), - destroyWorkspaceWebviews: vi.fn(), - isWebRuntimeSessionActive: vi.fn(() => true) - })) - -vi.mock('@/runtime/web-session-tabs-sync/actions/web-runtime-session-tab-lifecycle', () => ({ - closeWebRuntimeSessionTab -})) -vi.mock('@/runtime/hosted-terminals/creation/web-runtime-terminal-creation', () => ({ - createWebRuntimeSessionTerminal: vi.fn() -})) -vi.mock('@/runtime/web-session-tabs-sync/identity/web-runtime-session-environment', () => ({ - isWebRuntimeSessionActive -})) -vi.mock('@/store/slices/browser/cleanup/browser-webview-cleanup', () => ({ - destroyWorkspaceWebviews -})) - -const closeBrowserTab = vi.fn() -const closeUnifiedTab = vi.fn() -const recordClientHostedBrowserCloseIntents = vi.fn() - -function activeBrowserWorkspace(handle: { - environmentId?: string - staged?: true -}): Record { - return { - activeTabType: 'browser', - activeBrowserTabId: 'workspace-1', - activeWorktreeId: 'wt-1', - browserTabsByWorktree: { 'wt-1': [{ id: 'workspace-1' }] }, - browserPagesByWorkspace: { 'workspace-1': [{ id: 'page-1', workspaceId: 'workspace-1' }] }, - remoteBrowserPageHandlesByPageId: handle.environmentId - ? { - 'page-1': { - environmentId: handle.environmentId, - remotePageId: 'remote-1', - ...(handle.staged ? { staged: true } : {}) - } - } - : {}, - unifiedTabsByWorktree: { - 'wt-1': [{ id: 'unified-1', contentType: 'browser', entityId: 'workspace-1' }] - }, - closeBrowserTab, - closeUnifiedTab, - recordClientHostedBrowserCloseIntents - } -} - -/** Pageless host mirror ownership comes from its runtime-owned workspace row, not a local page. */ -function pagelessHostMirror(): Record { - return { - ...activeBrowserWorkspace({}), - browserPagesByWorkspace: {}, - worktreesByRepo: { - 'repo-1': [{ id: 'wt-1', repoId: 'repo-1', runtimeOwnerEnvironmentId: 'env-a' }] - } - } -} - -function requireListener(ref: { current: CloseActiveTabListener | null }): CloseActiveTabListener { - if (!ref.current) { - throw new Error('onCloseActiveTab was never registered') - } - return ref.current -} - -// Runtime connectivity alone does not prove host page ownership; inert host closes leave other tabs standing. -describe('useIpcEvents Close Tab on the active browser tab', () => { - beforeEach(() => { - vi.resetModules() - vi.unstubAllGlobals() - vi.clearAllMocks() - isWebRuntimeSessionActive.mockReturnValue(true) - }) - - // Staged pages lack host identity yet; inert remote close lets the in-flight snapshot recreate them. - it('unwinds a staged workspace as a cleanup close instead of closing it on the host', async () => { - const listenerRef: { current: CloseActiveTabListener | null } = { current: null } - await useIpcEventsForCloseRouting({ - closeActiveTabListenerRef: listenerRef, - getState: () => activeBrowserWorkspace({ environmentId: 'env-a', staged: true }) - }) - - requireListener(listenerRef)() - - expect(closeWebRuntimeSessionTab).not.toHaveBeenCalled() - expect(closeBrowserTab).toHaveBeenCalledWith('workspace-1', { reason: 'cleanup' }) - }) - - // Pageless mirrors need local removal by unified tab ID; host sync has no local page to retract. - it('removes the mirror tab itself when the host holds no page for it', async () => { - const listenerRef: { current: CloseActiveTabListener | null } = { current: null } - await useIpcEventsForCloseRouting({ - closeActiveTabListenerRef: listenerRef, - getState: pagelessHostMirror - }) - - requireListener(listenerRef)() - - expect(closeWebRuntimeSessionTab).toHaveBeenCalledWith({ - worktreeId: 'wt-1', - tabId: 'unified-1', - environmentId: 'env-a', - reason: 'user' - }) - expect(closeUnifiedTab).toHaveBeenCalledWith('unified-1', undefined) - expect(closeBrowserTab).not.toHaveBeenCalled() - }) - - it('tears a pageless mirror down here when its runtime is disconnected', async () => { - isWebRuntimeSessionActive.mockReturnValue(false) - const listenerRef: { current: CloseActiveTabListener | null } = { current: null } - await useIpcEventsForCloseRouting({ - closeActiveTabListenerRef: listenerRef, - getState: pagelessHostMirror - }) - - requireListener(listenerRef)() - - expect(closeWebRuntimeSessionTab).not.toHaveBeenCalled() - expect(closeUnifiedTab).toHaveBeenCalledWith('unified-1', undefined) - expect(closeBrowserTab).toHaveBeenCalledWith('workspace-1', undefined) - }) - - // Guest focus can leave the group’s active-type mirror stale; forwarded source IDs must still close the guest’s workspace. - it('closes the source workspace when the active-tab mirror points at a terminal', async () => { - const listenerRef: { current: CloseActiveTabListener | null } = { current: null } - await useIpcEventsForCloseRouting({ - closeActiveTabListenerRef: listenerRef, - getState: () => ({ - ...activeBrowserWorkspace({}), - activeTabType: 'terminal', - activeBrowserTabId: null, - browserPagesByWorkspace: { - 'workspace-1': [{ id: 'page-1', workspaceId: 'stale-workspace', worktreeId: 'stale-wt' }] - } - }) - }) - - requireListener(listenerRef)({ sourceId: 'page-1' }) - - expect(closeBrowserTab).toHaveBeenCalledWith('workspace-1', undefined) - expect(destroyWorkspaceWebviews).toHaveBeenCalled() - }) - - it('no-ops on a stale source id instead of closing the ambient active tab', async () => { - const listenerRef: { current: CloseActiveTabListener | null } = { current: null } - await useIpcEventsForCloseRouting({ - closeActiveTabListenerRef: listenerRef, - getState: () => activeBrowserWorkspace({}) - }) - - requireListener(listenerRef)({ sourceId: 'page-that-no-longer-exists' }) - - expect(closeBrowserTab).not.toHaveBeenCalled() - expect(closeUnifiedTab).not.toHaveBeenCalled() - expect(closeWebRuntimeSessionTab).not.toHaveBeenCalled() - }) - - it('no-ops when a source page belongs to no live workspace', async () => { - const listenerRef: { current: CloseActiveTabListener | null } = { current: null } - await useIpcEventsForCloseRouting({ - closeActiveTabListenerRef: listenerRef, - getState: () => ({ - ...activeBrowserWorkspace({}), - browserPagesByWorkspace: { - orphan: [{ id: 'orphan-page', workspaceId: 'workspace-1', worktreeId: 'wt-1' }] - } - }) - }) - - requireListener(listenerRef)({ sourceId: 'orphan-page' }) - - expect(closeBrowserTab).not.toHaveBeenCalled() - expect(closeUnifiedTab).not.toHaveBeenCalled() - expect(closeWebRuntimeSessionTab).not.toHaveBeenCalled() - }) -}) diff --git a/src/renderer/src/lib/workspace/creation/worktree-creation-backend-startup-focus.test.ts b/src/renderer/src/lib/workspace/creation/worktree-creation-backend-startup-focus.test.ts deleted file mode 100644 index 195a8b157ef..00000000000 --- a/src/renderer/src/lib/workspace/creation/worktree-creation-backend-startup-focus.test.ts +++ /dev/null @@ -1,189 +0,0 @@ -import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest' -import type * as MountedTerminalTabRegistryModule from '@/runtime/hosted-terminals/mounted-surfaces/mounted-terminal-tab-registry' -import type { PreloadApi } from '../../../../../preload/api-types' - -type CreateTerminalPayload = Parameters[0]>[0] - -const { focusRuntimeTerminalSurface } = vi.hoisted(() => ({ - focusRuntimeTerminalSurface: vi.fn(() => true) -})) - -vi.mock( - '@/runtime/hosted-terminals/mounted-surfaces/mounted-terminal-tab-registry', - async (importOriginal) => ({ - ...(await importOriginal()), - focusRuntimeTerminalSurface - }) -) - -import { useAppStore } from '@/store/index' -import { registerTerminalPresentationIpcBridge } from '@/hooks/ipc-events/terminal-presentation-ipc-bridge' -import { activateAndRevealWorktree } from '../activation/worktree-activation' -import { queueWorkspaceActivationTerminalFocus } from '../activation/workspace-activation-terminal-focus' -import { - makeCreatedAgentWorktree, - seedEmptyActivatableWorktree -} from '../activation/worktree-activation-created-agent.test-fixture' - -const initialAppStoreState = useAppStore.getState() -const AGENT_TAB_ID = 'b1a1c0de-0000-4000-8000-000000000001' -const AGENT_LEAF_ID = 'b1a1c0de-0000-4000-8000-000000000002' -const SETUP_TAB_ID = 'b1a1c0de-0000-4000-8000-000000000003' -const SETUP_LEAF_ID = 'b1a1c0de-0000-4000-8000-000000000004' - -let revealTerminal: (payload: CreateTerminalPayload) => void -let frames: FrameRequestCallback[] - -beforeEach(() => { - frames = [] - vi.stubGlobal('requestAnimationFrame', (callback: FrameRequestCallback) => { - frames.push(callback) - return frames.length - }) - vi.stubGlobal('cancelAnimationFrame', () => {}) - vi.stubGlobal( - 'MutationObserver', - class { - observe(): void {} - disconnect(): void {} - } - ) - vi.stubGlobal('document', { - activeElement: null, - body: {}, - querySelector: () => null, - addEventListener: () => {}, - removeEventListener: () => {} - }) - vi.stubGlobal('window', { - api: { - ui: { - onCreateTerminal: (listener: typeof revealTerminal) => { - revealTerminal = listener - return () => {} - }, - onLaunchRecipe: () => () => {}, - onRequestTerminalTabMount: () => () => {}, - replyTerminalCreate: vi.fn() - } - } - }) - registerTerminalPresentationIpcBridge([]) -}) - -afterEach(() => { - vi.unstubAllGlobals() - useAppStore.setState(initialAppStoreState, true) -}) - -describe('a local agent create watched to completion', () => { - it('opens on and focuses the agent tab the host adopted without activating', () => { - const worktree = makeCreatedAgentWorktree() - seedEmptyActivatableWorktree(worktree) - useAppStore.setState({ activeWorktreeId: null, activeTabId: null }) - - // The host reveals its startup and setup terminals as main sends them for a local create. - revealTerminal({ - requestId: 'reveal-agent', - worktreeId: worktree.id, - ptyId: 'pty-agent', - launchAgent: 'codex', - activate: false, - surfaceOwner: false, - tabId: AGENT_TAB_ID, - leafId: AGENT_LEAF_ID - }) - revealTerminal({ - requestId: 'reveal-setup', - worktreeId: worktree.id, - ptyId: 'pty-setup', - title: 'Setup', - activate: false, - surfaceOwner: false, - tabId: SETUP_TAB_ID, - leafId: SETUP_LEAF_ID - }) - expect(useAppStore.getState().activeWorktreeId).toBeNull() - - // The completion rule's activation for a user still on the creation panel. - const activation = activateAndRevealWorktree(worktree.id, { - sidebarRevealBehavior: 'auto', - agent: 'codex', - backendStartupTerminalSpawned: true - }) - expect(activation).toEqual({ primaryTabId: null }) - expect(useAppStore.getState().tabsByWorktree[worktree.id]?.map((tab) => tab.id)).toEqual([ - AGENT_TAB_ID, - SETUP_TAB_ID - ]) - - expect(queueWorkspaceActivationTerminalFocus(worktree.id, activation)).toBe(true) - frames.shift()?.(0) - expect(focusRuntimeTerminalSurface).toHaveBeenCalledWith(AGENT_TAB_ID, null, worktree.id) - }) -}) - -describe('host reveal after early launch publication fails', () => { - it.each([ - { viewer: 'focus-in-workspace', moved: true, follows: false }, - { viewer: 'focus-in-workspace', moved: false, follows: true }, - { viewer: 'focus-window', moved: true, follows: true }, - { viewer: 'none', moved: true, follows: false } - ] as const)( - 'honors $viewer after navigation=$moved for new and existing panes', - ({ viewer, moved, follows }) => { - const worktree = makeCreatedAgentWorktree() - const other = { ...worktree, id: `${worktree.id}-other` } - const { revealWorktreeInSidebar } = seedEmptyActivatableWorktree(worktree, { - extraWorktrees: [other] - }) - const activeWorkspace = moved ? other.id : worktree.id - useAppStore.setState({ activeWorktreeId: activeWorkspace, activeTabId: null }) - focusRuntimeTerminalSurface.mockClear() - const payload: CreateTerminalPayload = { - requestId: 'reveal-agent', - worktreeId: worktree.id, - ptyId: 'pty-agent', - launchAgent: 'codex', - activate: false, - surfaceOwner: false, - agentLaunchViewer: viewer, - tabId: AGENT_TAB_ID, - leafId: AGENT_LEAF_ID - } - revealTerminal(payload) - expect(useAppStore.getState().activeWorktreeId).toBe(follows ? worktree.id : activeWorkspace) - expect(useAppStore.getState().tabsByWorktree[worktree.id]?.map((tab) => tab.id)).toEqual([ - AGENT_TAB_ID - ]) - expect(useAppStore.getState().terminalLayoutsByTabId[AGENT_TAB_ID]?.ptyIdsByLeafId).toEqual({ - [AGENT_LEAF_ID]: 'pty-agent' - }) - if (follows) { - expect(useAppStore.getState().activeTabId).toBe(AGENT_TAB_ID) - expect(focusRuntimeTerminalSurface).toHaveBeenCalledWith( - AGENT_TAB_ID, - AGENT_LEAF_ID, - worktree.id - ) - } else { - expect(focusRuntimeTerminalSurface).not.toHaveBeenCalled() - expect(useAppStore.getState().activeTabId).toBeNull() - } - if (viewer === 'focus-in-workspace') { - expect(revealWorktreeInSidebar).not.toHaveBeenCalled() - } - // A reveal can find a tab restored while its earlier publication reply was lost. - useAppStore.setState({ activeWorktreeId: activeWorkspace, activeTabId: null }) - focusRuntimeTerminalSurface.mockClear() - revealTerminal(payload) - expect(useAppStore.getState().activeWorktreeId).toBe(follows ? worktree.id : activeWorkspace) - expect(useAppStore.getState().tabsByWorktree[worktree.id]).toHaveLength(1) - if (follows) { - expect(focusRuntimeTerminalSurface).toHaveBeenCalledOnce() - } else { - expect(focusRuntimeTerminalSurface).not.toHaveBeenCalled() - } - } - ) -}) diff --git a/src/renderer/src/runtime/files/reads/runtime-paged-file-read.test.ts b/src/renderer/src/runtime/files/reads/runtime-paged-file-read.test.ts index 4fb03894d76..ddf50bf894f 100644 --- a/src/renderer/src/runtime/files/reads/runtime-paged-file-read.test.ts +++ b/src/renderer/src/runtime/files/reads/runtime-paged-file-read.test.ts @@ -1,13 +1,10 @@ import { describe, expect, it } from 'vitest' import { readRuntimeFileContent } from './runtime-file-read-client' import { - fsReadFileChunk, - fsStat, installRuntimeFileClientEnvironment, runtimeEnvironmentCall, type RuntimeRpcRequest } from '../runtime-file-client-test-harness' -import { readRuntimeFileBytes } from './runtime-paged-file-read' installRuntimeFileClientEnvironment() @@ -136,16 +133,6 @@ describe('remote reads larger than one RPC reply', () => { expect(Buffer.from(result.content, 'utf8').equals(bytes)).toBe(true) }) - it('keeps a UTF-8 BOM so a save does not drop it', async () => { - const bytes = Buffer.concat([Buffer.from([0xef, 0xbb, 0xbf]), htmlBytes(MB)]) - serveRemoteFile('table.csv', { bytes }) - - const result = await readRuntimeFileContent(remote('table.csv')) - - expect(result.content.charCodeAt(0)).toBe(0xfeff) - expect(Buffer.from(result.content, 'utf8').equals(bytes)).toBe(true) - }) - it.each([2, 5])('opens a %i MB PDF in the editor', async (sizeMb) => { const bytes = patternedBytes(sizeMb * MB) serveRemoteFile('paper.pdf', { bytes }) @@ -159,17 +146,6 @@ describe('remote reads larger than one RPC reply', () => { expect(Buffer.from(result.content, 'base64').equals(bytes)).toBe(true) }) - it('treats a large file with a NUL byte as binary, like a local read', async () => { - const bytes = htmlBytes(2 * MB) - bytes[10] = 0 - serveRemoteFile('data.txt', { bytes }) - - await expect(readRuntimeFileContent(remote('data.txt'))).resolves.toEqual({ - content: '', - isBinary: true - }) - }) - it('rejects a chunk whose payload does not match its byte count', async () => { serveRemoteFile('report.html', { bytes: htmlBytes(2 * MB), @@ -181,30 +157,3 @@ describe('remote reads larger than one RPC reply', () => { ) }) }) - -describe('whole-file bytes for a bytes custom editor', () => { - it('returns the exact bytes of a paged remote workbook', async () => { - const bytes = patternedBytes(MB + 17) - serveRemoteFile('book.xlsx', { bytes }) - - const result = await readRuntimeFileBytes(remote('book.xlsx'), 16 * MB) - - expect(result.kind).toBe('ok') - expect(result.kind === 'ok' && Buffer.from(result.bytes).equals(bytes)).toBe(true) - }) - - it('reads a local file through the local stat and chunk IPC', async () => { - const bytes = patternedBytes(300 * 1024) - fsStat.mockResolvedValue({ size: bytes.length, isDirectory: false, mtime: 5 }) - fsReadFileChunk.mockImplementation((args: ChunkParams) => - Promise.resolve(sliceChunk(args, bytes)) - ) - - const result = await readRuntimeFileBytes( - { target: { kind: 'local' }, filePath: '/repo/book.xlsx' }, - 16 * MB - ) - - expect(result.kind === 'ok' && Buffer.from(result.bytes).equals(bytes)).toBe(true) - }) -}) diff --git a/src/renderer/src/store/repos/repos.test.ts b/src/renderer/src/store/repos/repos.test.ts index a75c47051a8..0908462a5de 100644 --- a/src/renderer/src/store/repos/repos.test.ts +++ b/src/renderer/src/store/repos/repos.test.ts @@ -2,22 +2,16 @@ import { describe, expect, it, vi } from 'vitest' import { getRepoExecutionHostId } from '../../../../shared/execution-host' import { createTestStore } from '../testing/app-store-test-fixture' import { makeWorktree } from '../slices/worktrees/worktree.test-fixture' -import { workItemsCacheKey } from '../github/cache/cache-identity' import type { Repo } from '../../../../shared/repo-types' -import { toast } from 'sonner' import '../../../../shared/global-settings-test-fixture' import '../../web/preload-api/web-fallback-api' import { installReposRuntimeRoutingHarness, - localRepo, - projectsSetupExistingFolder, ptyKill, remoteRepo, - reposRemoveForHost, runtimeEnvironmentCall, - runtimeEnvironmentsList, - sshRepo + runtimeEnvironmentsList } from './repo-runtime-routing.test-fixture' vi.mock('sonner', () => ({ @@ -36,81 +30,6 @@ const ownerOf = (repo: Repo) => ({ hostId: getRepoExecutionHostId(repo) }) const serverRepo: Repo = { ...remoteRepo, executionHostId: 'runtime:env-1' } describe('repo slice runtime routing', () => { - it('names an SSH host without a server and offers Connect, never its raw id (#37767)', async () => { - const store = createTestStore() - store.setState({ sshTargetLabels: new Map([['ssh-1', 'build box']]) }) - const connect = vi.fn().mockResolvedValue(null) - Object.assign(window.api, { ssh: { connect } }) - - await expect( - store.getState().setupProjectExistingFolder({ - projectId: 'github:acme/orca', - hostId: 'ssh:ssh-1', - path: '/srv/orca', - kind: 'git' - }) - ).resolves.toBeNull() - - expect(projectsSetupExistingFolder).not.toHaveBeenCalled() - const [, options] = vi.mocked(toast.error).mock.calls[0] ?? [] - expect(options?.description).toContain('build box') - expect(options?.description).not.toContain('ssh-1') - const action = options?.action - if (!action || typeof action !== 'object' || !('onClick' in action)) { - throw new Error('expected a Connect action') - } - // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: the action ignores its click event. - action.onClick({} as never) - expect(connect).toHaveBeenCalledWith({ targetId: 'ssh-1' }) - }) - - it('drops persisted visit timestamps for removed unhydrated SSH repos', async () => { - const store = createTestStore() - const sshWorktreeId = `${sshRepo.id}::/home/orca/wt` - const localWorktreeId = `${localRepo.id}::/local/wt` - store.setState({ - repos: [sshRepo, localRepo], - activeRepoId: sshRepo.id, - lastVisitedAtByWorktreeId: { - [sshWorktreeId]: 100, - [localWorktreeId]: 200 - } - }) - - await store.getState().removeProject(sshRepo.id, ownerOf(sshRepo)) - - expect(store.getState().repos).toEqual([localRepo]) - expect(store.getState().lastVisitedAtByWorktreeId).toEqual({ [localWorktreeId]: 200 }) - expect(reposRemoveForHost).toHaveBeenCalledWith({ repoId: sshRepo.id, hostId: 'ssh:ssh-1' }) - }) - - it('evicts GitHub caches for removed repos using repo id and legacy path keys', async () => { - const store = createTestStore() - store.setState({ - repos: [localRepo], - workItemsInvalidationNonce: 2, - workItemsCache: { - [workItemsCacheKey(localRepo.id, 20, '')]: { data: [], fetchedAt: 1 }, - [workItemsCacheKey(localRepo.path, 20, '')]: { data: [], fetchedAt: 1 }, - [workItemsCacheKey('other-repo', 20, '')]: { data: [], fetchedAt: 1 } - }, - prCache: { - [`${localRepo.id}::branch`]: { data: {} as never, fetchedAt: 1 }, - [`${localRepo.path}::branch`]: { data: {} as never, fetchedAt: 1 }, - 'other-repo::branch': { data: {} as never, fetchedAt: 1 } - } - }) - - await store.getState().removeProject(localRepo.id, ownerOf(localRepo)) - - expect(Object.keys(store.getState().workItemsCache)).toEqual([ - workItemsCacheKey('other-repo', 20, '') - ]) - expect(Object.keys(store.getState().prCache)).toEqual(['other-repo::branch']) - // A refetch signal only: eviction and the retired source owner may each bump it. - expect(store.getState().workItemsInvalidationNonce).toBeGreaterThan(2) - }) - it('stops remote runtime terminals instead of killing remote ids through local pty IPC', async () => { runtimeEnvironmentsList.mockResolvedValue([{ id: 'env-1' }]) runtimeEnvironmentCall.mockResolvedValue({ @@ -148,46 +67,4 @@ describe('repo slice runtime routing', () => { expect(ptyKill).toHaveBeenCalledWith('pty-local-stale') expect(ptyKill).not.toHaveBeenCalledWith('remote:term-1') }) - - it('cleans up hidden detected worktree state when removing a repo', async () => { - const store = createTestStore() - const hiddenWorktree = makeWorktree({ - id: `${localRepo.id}::/local/hidden`, - repoId: localRepo.id, - path: '/local/hidden' - }) - store.setState({ - repos: [localRepo], - worktreesByRepo: { [localRepo.id]: [] }, - detectedWorktreesByRepo: { - [localRepo.id]: { - repoId: localRepo.id, - authoritative: true, - source: 'git', - worktrees: [ - { - ...hiddenWorktree, - ownership: 'external', - selectedCheckout: false, - visible: false - } - ] - } - }, - tabsByWorktree: { - [hiddenWorktree.id]: [{ id: 'tab-hidden', worktreeId: hiddenWorktree.id }] as never - }, - ptyIdsByTabId: { - 'tab-hidden': ['pty-hidden'] - }, - activeWorktreeId: hiddenWorktree.id - }) - - await store.getState().removeProject(localRepo.id, ownerOf(localRepo)) - - expect(store.getState().detectedWorktreesByRepo[localRepo.id]).toBeUndefined() - expect(store.getState().tabsByWorktree[hiddenWorktree.id]).toBeUndefined() - expect(store.getState().activeWorktreeId).toBeNull() - expect(ptyKill).toHaveBeenCalledWith('pty-hidden') - }) }) diff --git a/src/renderer/src/store/slices/browser/hosts/browser-remote-tab-creation.test.ts b/src/renderer/src/store/slices/browser/hosts/browser-remote-tab-creation.test.ts deleted file mode 100644 index b600a0c6624..00000000000 --- a/src/renderer/src/store/slices/browser/hosts/browser-remote-tab-creation.test.ts +++ /dev/null @@ -1,197 +0,0 @@ -import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest' -import type { AppState } from '../../../types' -import { layoutApplyTestTransport } from '@/runtime/workspace-layout/testing/layout-apply-test-transport.test-fixture' -import { - createBrowserMockApi, - resetBrowserRuntimeMocks, - runtimeStatuses, - settingsWithRuntime -} from '../browser-slice-test-harness' -import { - createTestStoreDrawnFromMain, - onePaneGroup -} from '../browser-slice-drawn-from-main.test-fixture' - -const createWebRuntimeSessionBrowserTabMock = vi.hoisted(() => vi.fn()) -const runtimeEnvironmentCall = vi.fn() -const runtimeEnvironmentTransportCall = vi.fn() - -vi.mock('@/runtime/rpc/calls/runtime-rpc-call', async (importOriginal) => - ( - await import('@/runtime/workspace-layout/testing/layout-apply-test-transport.test-fixture') - ).runtimeRpcClientWithLayoutApply(await importOriginal()) -) -vi.mock('@/runtime/workspace-layout/frames/runtime-layout-client', async (importOriginal) => - ( - await import('@/runtime/workspace-layout/testing/layout-apply-test-transport.test-fixture') - ).runtimeLayoutClientWithTestFrames(await importOriginal()) -) -vi.mock('@/store/slices/browser/tabs/paired-browser-tab-creator', () => ({ - getRegisteredPairedBrowserTabCreator: () => null, - loadPairedBrowserTabCreator: async () => createWebRuntimeSessionBrowserTabMock -})) - -const mockApi = createBrowserMockApi(runtimeEnvironmentTransportCall) - -// @ts-expect-error test window mock -globalThis.window = { api: mockApi } - -/** A server-owned workspace's desktop tab: the window draws it, as main never holds that layout. */ -function expectDrawnHere(state: AppState): void { - expect(layoutApplyTestTransport.sent).toEqual([]) - expect(state.createUnifiedTab).toHaveBeenCalled() -} - -describe('createBrowserSlice runtime guard', () => { - beforeEach(() => { - resetBrowserRuntimeMocks({ - runtimeEnvironmentCall, - runtimeEnvironmentTransportCall, - createWebRuntimeSessionBrowserTabMock - }) - }) - - afterEach(() => { - vi.unstubAllGlobals() - }) - - it('rejects direct client-local materialization in the web client', async () => { - vi.stubGlobal('__ORCA_WEB_CLIENT__', true) - const { store } = await createTestStoreDrawnFromMain() - - expect(() => store.getState().createBrowserTab('wt-1', 'about:blank')).toThrow( - 'must be created by a capable paired runtime' - ) - expect(store.getState().browserTabsByWorktree['wt-1']).toBeUndefined() - expect(store.getState().createUnifiedTab).not.toHaveBeenCalled() - }) - - it('uses the desktop client browser when a remote npm host cannot stream', async () => { - const { store, main } = await createTestStoreDrawnFromMain({ - ...onePaneGroup('wt-remote'), - activeWorktreeId: 'wt-remote', - settings: settingsWithRuntime('env-1'), - runtimeStatusByEnvironmentId: runtimeStatuses([]), - worktreesByRepo: { - 'repo-1': [ - { - id: 'wt-remote', - repoId: 'repo-1', - hostId: 'local', - runtimeOwnerEnvironmentId: 'env-1' - } as never - ] - } - }) - - await store.getState().openNewBrowserTabInActiveWorkspace('group-1') - await main.settle() - - expect(createWebRuntimeSessionBrowserTabMock).not.toHaveBeenCalled() - const tab = store.getState().browserTabsByWorktree['wt-remote']?.[0] - const page = tab ? store.getState().browserPagesByWorkspace[tab.id]?.[0] : undefined - expect(page?.browserRuntimeEnvironmentId).toBeNull() - expectDrawnHere(store.getState()) - }) - - it('does not create a local fallback tab when remote browser creation fails', async () => { - // Failed remote creation must not fall back local and split workspace ownership (#5321). - createWebRuntimeSessionBrowserTabMock.mockResolvedValueOnce(false) - const { store, main } = await createTestStoreDrawnFromMain({ - ...onePaneGroup('wt-remote'), - activeWorktreeId: 'wt-remote', - settings: { activeRuntimeEnvironmentId: 'env-1' } as AppState['settings'], - runtimeStatusByEnvironmentId: runtimeStatuses(['browser.screencast.v1']), - worktreesByRepo: { - 'repo-1': [ - { - id: 'wt-remote', - repoId: 'repo-1', - hostId: 'local', - runtimeOwnerEnvironmentId: 'env-1' - } as never - ] - } - }) - - await expect(store.getState().openNewBrowserTabInActiveWorkspace('group-1')).rejects.toThrow( - 'The paired runtime could not create a managed browser tab.' - ) - - expect(createWebRuntimeSessionBrowserTabMock).toHaveBeenCalledWith({ - worktreeId: 'wt-remote', - environmentId: 'env-1', - url: 'about:blank', - targetGroupId: 'group-1', - clientTargetGroupId: 'group-1' - }) - // No local tab created, no unified tab, no feature interaction recorded. - await main.settle() - expect(layoutApplyTestTransport.sent).toEqual([]) - expect(store.getState().browserTabsByWorktree['wt-remote']).toBeUndefined() - expect(store.getState().createUnifiedTab).not.toHaveBeenCalled() - expect(store.getState().recordFeatureInteraction).not.toHaveBeenCalledWith( - 'browser-tab-created' - ) - }) - - it('rejects a paired-web sign-in tab when the remote host cannot stream browsers', async () => { - vi.stubGlobal('__ORCA_WEB_CLIENT__', true) - const { store, main } = await createTestStoreDrawnFromMain({ - ...onePaneGroup('wt-remote'), - activeWorktreeId: 'wt-remote', - settings: settingsWithRuntime('env-1'), - runtimeStatusByEnvironmentId: runtimeStatuses([]), - worktreesByRepo: { - 'repo-1': [ - { - id: 'wt-remote', - repoId: 'repo-1', - hostId: 'local', - runtimeOwnerEnvironmentId: 'env-1' - } as never - ] - } - }) - - await expect( - store - .getState() - .openBrowserProfileTabInActiveWorkspace('https://accounts.google.com/', 'profile-1') - ).resolves.toBe(false) - - expect(createWebRuntimeSessionBrowserTabMock).not.toHaveBeenCalled() - expect(store.getState().browserTabsByWorktree['wt-remote']).toBeUndefined() - await main.settle() - expect(layoutApplyTestTransport.sent).toEqual([]) - }) - - it('does not create a local fallback tab when remote browser creation throws', async () => { - createWebRuntimeSessionBrowserTabMock.mockRejectedValueOnce(new Error('remote down')) - const { store, main } = await createTestStoreDrawnFromMain({ - ...onePaneGroup('wt-remote'), - activeWorktreeId: 'wt-remote', - settings: { activeRuntimeEnvironmentId: 'env-1' } as AppState['settings'], - runtimeStatusByEnvironmentId: runtimeStatuses(['browser.screencast.v1']), - worktreesByRepo: { - 'repo-1': [ - { - id: 'wt-remote', - repoId: 'repo-1', - hostId: 'local', - runtimeOwnerEnvironmentId: 'env-1' - } as never - ] - } - }) - - await expect(store.getState().openNewBrowserTabInActiveWorkspace('group-1')).rejects.toThrow( - 'remote down' - ) - await main.settle() - expect(layoutApplyTestTransport.sent).toEqual([]) - - expect(store.getState().browserTabsByWorktree['wt-remote']).toBeUndefined() - expect(store.getState().createUnifiedTab).not.toHaveBeenCalled() - }) -}) diff --git a/src/renderer/src/store/slices/worktrees/session/folder-workspace-activation-and-activity.test.ts b/src/renderer/src/store/slices/worktrees/session/folder-workspace-activation-and-activity.test.ts index 34d13c9df85..be92d9d6b35 100644 --- a/src/renderer/src/store/slices/worktrees/session/folder-workspace-activation-and-activity.test.ts +++ b/src/renderer/src/store/slices/worktrees/session/folder-workspace-activation-and-activity.test.ts @@ -1,5 +1,5 @@ import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest' -import { installAttentionScreen } from '../../workspace-attention/workspace-unread-characterization-fixture' +import '../../workspace-attention/workspace-unread-characterization-fixture' import type { AppState } from '../../../types' import type { FolderWorkspace } from '../../../../../../shared/folder-workspaces/folder-workspace-types' import type { ProjectGroup } from '../../../../../../shared/project-group-types' @@ -101,28 +101,6 @@ afterEach(() => { }) describe('folder workspace generic activation and activity', () => { - it('publishes folder-scoped active state and reads the flag once on screen', async () => { - const folderWorkspace = makeFolderWorkspace({ isUnread: true }) - const workspaceKey = folderWorkspaceKey(folderWorkspace.id) - folderWorkspacesUpdate.mockImplementation(respondWithUpdates(folderWorkspace)) - const store = seedLocalFolderStore(folderWorkspace) - - installAttentionScreen(store, true) - store.getState().setActiveWorktree(workspaceKey) - await Promise.resolve() - - expect(store.getState()).toMatchObject({ - activeRepoId: null, - activeWorktreeId: workspaceKey, - activeWorkspaceKey: workspaceKey - }) - expect(store.getState().folderWorkspaces[0]?.isUnread).toBe(false) - expect(folderWorkspacesUpdate).toHaveBeenCalledWith({ - folderWorkspaceId: folderWorkspace.id, - updates: { isUnread: false } - }) - }) - it('coalesces repeated activity persistence while keeping local activity current', async () => { vi.useFakeTimers() vi.setSystemTime(1_000) From c20f528d3e7493642b7701bd366f3e5da6551bf2 Mon Sep 17 00:00:00 2001 From: Kelvin Amoaba <97001695+AmoabaKelvin@users.noreply.github.com> Date: Wed, 11 Nov 2026 00:27:08 +0000 Subject: [PATCH 03/21] fix(omp): show the picked model's own thinking levels before the first chat message (#40933) OMP's thinking menu belongs to the running model, but Orca copied it onto every other OMP model in a new chat. Treat OMP like OpenCode: leave non-running models' menus unknown, drop such copies saved by older builds, and read a picked model's menu from a throwaway listing session started on that model. Starting on it, rather than switching to it, keeps OMP from recording the model as used in the user's account; if OMP cannot start on it, the listing runs again without it. A read that asks for a picked model's menu now reports the listing in progress, so an open picker updates when the menu lands instead of on the next open. Fixes #39015 --- .../discovery/acp-listing-session-models.ts | 148 ++++++++++++ .../discovery/acp-model-catalog-probe.test.ts | 107 +++++++++ .../acp/discovery/acp-model-catalog-probe.ts | 216 +++++------------- .../fixtures/omp-v18.17.0-model-efforts.json | 102 +++++++++ .../acp/options/acp-structured-options.ts | 5 +- ...del-catalog-effort-target-progress.test.ts | 50 ++++ .../agent-model-catalog-entry.test.ts | 16 +- .../agent-model-catalog-models.ts | 4 +- .../agent-model-catalog-snapshot.ts | 5 +- .../options/agent-model-catalog-ownership.ts | 2 +- 10 files changed, 481 insertions(+), 174 deletions(-) create mode 100644 src/main/acp/discovery/acp-listing-session-models.ts create mode 100644 src/main/acp/fixtures/omp-v18.17.0-model-efforts.json create mode 100644 src/main/native-chat/agent-model-catalog/agent-model-catalog-effort-target-progress.test.ts diff --git a/src/main/acp/discovery/acp-listing-session-models.ts b/src/main/acp/discovery/acp-listing-session-models.ts new file mode 100644 index 00000000000..0c35c338adc --- /dev/null +++ b/src/main/acp/discovery/acp-listing-session-models.ts @@ -0,0 +1,148 @@ +import { setTimeout as delay } from 'node:timers/promises' +import { observeAcpForkSupport } from '../capabilities/acp-fork-support' +import { + assertStructuredAgentResolutionCurrent, + type StructuredAgentResolutionOperation +} from '../../native-chat/structured-agent-command-resolution' +import type { AgentSessionModelOption } from '../../../shared/agent-session/agent-session-wire' +import type { AcpSessionModelDiscovery } from './acp-model-discovery-policy' +import type { AcpLaunchSpec } from '../launch/acp-launch-contract' +import type { AcpSessionStarted, AcpSessionStartOptions } from '../protocol/acp-session-setup' +import { AcpStructuredOptions } from '../options/acp-structured-options' +import type { + InitializeResponse, + SetSessionConfigOptionResponse +} from '../generated/acp-protocol.generated' +import { withTimeout } from '../../../shared/promise-timeout-fallback' +import { AcpAgentError } from '../transport/acp-errors' + +/** Session-free capability boundary: no session methods are exposed. */ +export type AcpModelCatalogConnection = { + initialize(): Promise + requestSessionFreeExtension(method: string, params: unknown): Promise + close(): Promise +} + +export type AcpModelCatalogSessionConnection = Omit< + AcpModelCatalogConnection, + 'requestSessionFreeExtension' +> & { + start(options: AcpSessionStartOptions): Promise + setConfigOption(configId: string, value: string): Promise + deleteSession(sessionId: string): Promise +} + +// OpenCode can answer -32603 while its providers are still opening; a later attempt lists them. +async function startListingSession( + connection: AcpModelCatalogSessionConnection, + workspace: string, + operation?: StructuredAgentResolutionOperation +): Promise { + for (const backoffMs of [250, 500]) { + try { + return await connection.start({ cwd: workspace, mcpServers: [] }) + } catch (error) { + if (!(error instanceof AcpAgentError && error.code === -32603)) { + throw error + } + } + await delay(backoffMs, undefined, { signal: operation?.signal }) + assertStructuredAgentResolutionCurrent(operation) + } + return connection.start({ cwd: workspace, mcpServers: [] }) +} + +/** Lists models from a disposable session, learning menus for the models it starts on or switches to. */ +export async function readSessionModels( + spec: AcpLaunchSpec, + listing: AcpSessionModelDiscovery, + connection: AcpModelCatalogSessionConnection, + workspace: string, + operation?: StructuredAgentResolutionOperation, + selectedModels?: readonly string[], + launchedModel?: string, + ready?: (cleanup: () => Promise) => void +): Promise<{ models: AgentSessionModelOption[]; attemptedModels: string[] }> { + const initialized = await connection.initialize() + assertStructuredAgentResolutionCurrent(operation) + observeAcpForkSupport(spec.agent, initialized) + const deletes = Boolean(initialized.agentCapabilities?.sessionCapabilities?.delete) + if (!deletes && !listing.sessionFolderArgs) { + throw new Error(`${spec.agent} cannot delete a session, so it lists its models only in chats`) + } + const started = await startListingSession(connection, workspace, operation) + let stopped = false + const cleanup = async (): Promise => { + stopped = true + if (deletes) { + await connection.deleteSession(started.sessionId).catch((error: unknown) => { + console.warn(`[acp-model-catalog] ${spec.agent} kept its listing session:`, error) + }) + } + } + // Startup and menu discovery have separate budgets; cleanup also runs on cancellation. + ready?.(cleanup) + const options = new AcpStructuredOptions(spec.dialect) + options.adoptSession(started.response, 'new') + let models = options.read().models + const attemptedModels: string[] = [] + try { + if (launchedModel !== undefined) { + attemptedModels.push(launchedModel) + const running = options.read().current.model + // An unresolved --model leaves OMP on its default with a menu that is not that model's. + if (running !== launchedModel) { + models = models.map((row) => + row.id === running ? { ...row, efforts: [], effortsKnown: false as const } : row + ) + } + } else if (spec.dialect.effortChangesModel) { + const byId = new Map(models.map((model) => [model.id, model])) + const targets = selectedModels ? selectedModels.flatMap((id) => byId.get(id) ?? []) : models + const sweep = async () => { + for (const model of targets) { + if (stopped) { + return + } + assertStructuredAgentResolutionCurrent(operation) + if (model.effortsKnown !== false) { + continue + } + attemptedModels.push(model.id) + try { + for (let attempt = 0; attempt < 2 && !stopped; attempt++) { + const write = options.write('model', model.id) + if (write?.method !== 'config') { + break + } + const reply = await connection.setConfigOption(write.configId, write.value) + if (stopped) { + return + } + assertStructuredAgentResolutionCurrent(operation) + options.adoptConfigOptions(reply.configOptions) + const observed = options.read() + if (observed.current.model !== model.id) { + continue + } + const known = observed.models.find((row) => row.id === model.id) + if (known) { + models = models.map((row) => (row.id === model.id ? known : row)) + } + break + } + } catch { + assertStructuredAgentResolutionCurrent(operation) + } + } + } + await withTimeout(sweep(), 5_000, undefined) + } + assertStructuredAgentResolutionCurrent(operation) + return { models, attemptedModels } + } finally { + if (!ready) { + await cleanup() + } + } +} diff --git a/src/main/acp/discovery/acp-model-catalog-probe.test.ts b/src/main/acp/discovery/acp-model-catalog-probe.test.ts index bbb5943d4b2..85f040a7fb3 100644 --- a/src/main/acp/discovery/acp-model-catalog-probe.test.ts +++ b/src/main/acp/discovery/acp-model-catalog-probe.test.ts @@ -1,3 +1,6 @@ +import { mkdtemp, rm } from 'node:fs/promises' +import { tmpdir } from 'node:os' +import { join } from 'node:path' import { afterEach, describe, expect, it } from 'vitest' import { acpLaunchSpecFor } from '../launch/acp-launch-registry' import type { AcpLaunchSpec } from '../launch/acp-launch-contract' @@ -5,6 +8,10 @@ import { createAcpModelCatalogProbe } from './acp-model-catalog-probe' import { AcpSessionRuntime } from '../protocol/acp-session-runtime' import { AcpScriptedAgent } from '../process/testing/acp-scripted-agent.test-support' import { openCodeAcpAccountBinding } from '../../opencode/opencode-structured-account-home' +import { AcpAgentExitedError } from '../transport/acp-errors' +// Real `omp acp` 18.17.0 session/new replies: started on its default fake/wide, with +// `--model fake/narrow`, and with `--model fake/gone`, which it cannot resolve. +import recorded from '../fixtures/omp-v18.17.0-model-efforts.json' // Bypass the app's managed-account service for this fixture. const OPENCODE: AcpLaunchSpec = { @@ -87,3 +94,103 @@ describe('ACP model catalog probes', () => { ]) }) }) + +async function listOmp(selectedModels?: string[]) { + const agent = new AcpScriptedAgent() + agents.push(agent) + let launchedModel: string | undefined + agent.on('initialize', (frame) => agent.reply(frame, { protocolVersion: 1 })) + agent.on('session/new', (frame) => + agent.reply(frame, { + sessionId: 'probe-session', + ...(launchedModel === 'fake/narrow' + ? recorded.narrow + : launchedModel === 'fake/gone' + ? recorded.unresolved + : recorded.new) + }) + ) + const dataDirectory = await mkdtemp(join(tmpdir(), 'orca-omp-efforts-')) + try { + const listing = await createAcpModelCatalogProbe(acpLaunchSpecFor('omp')!, { + resolveEnvironment: async () => ({ PATH: '/usr/bin', HOME: '/home/user' }), + resolveCommand: () => '/opt/omp/bin/omp', + probeVersion: async () => ({ version: '18.17.0' }), + homePath: '/home/user', + dataDirectory, + connectSession: (launch) => { + const at = launch.args.indexOf('--model') + launchedModel = at === -1 ? undefined : launch.args[at + 1] + if (launchedModel?.includes(':')) { + // OMP 18.17.0 prints `Model "fake/gone:high" not found.` and exits 1 before ACP starts. + const exited = async () => { + throw new AcpAgentExitedError('Model "fake/gone:high" not found.') + } + return { + initialize: exited, + start: exited, + setConfigOption: exited, + deleteSession: exited, + close: async () => {} + } + } + const runtime = new AcpSessionRuntime(agent.stdout, agent.stdin) + return { + initialize: () => runtime.initialize(), + start: (options) => runtime.start(options), + setConfigOption: (id, value) => runtime.setConfigOption(id, value), + deleteSession: (sessionId) => runtime.deleteSession(sessionId), + close: async () => runtime.close() + } + } + })( + { variable: 'PI_CODING_AGENT_DIR', path: '/homes/omp' }, + { workspacePath: '/work/project', ...(selectedModels ? { selectedModels } : {}) } + ) + // A model switch would record usage in the user's OMP account. + expect(agent.frames.map((frame) => frame.method)).toEqual(['initialize', 'session/new']) + return (id: string) => listing.models.find((model) => model.id === id) + } finally { + await rm(dataDirectory, { recursive: true, force: true }) + } +} + +const levels = (model: { efforts: { value: string }[] } | undefined) => + model?.efforts.map((effort) => effort.value) + +it("lists a picked OMP model with its own thinking levels, not the startup model's", async () => { + const listed = await listOmp() + expect(levels(listed('fake/wide'))).toEqual([ + 'off', + 'auto', + 'minimal', + 'low', + 'medium', + 'high', + 'xhigh' + ]) + expect(listed('fake/narrow')).toMatchObject({ efforts: [], effortsKnown: false }) + + const picked = await listOmp(['fake/narrow']) + expect(levels(picked('fake/narrow'))).toEqual(['off', 'auto', 'low', 'high']) + expect(picked('fake/narrow')).not.toHaveProperty('effortsKnown') + expect(picked('fake/wide')).toMatchObject({ efforts: [], effortsKnown: false }) +}) + +it('trusts no menu when OMP cannot start on the picked model', async () => { + const picked = await listOmp(['fake/gone']) + expect(picked('fake/wide')).toMatchObject({ efforts: [], effortsKnown: false }) +}) + +it('still lists OMP when it cannot start on the picked model', async () => { + const picked = await listOmp(['fake/gone:high']) + expect(levels(picked('fake/wide'))).toEqual([ + 'off', + 'auto', + 'minimal', + 'low', + 'medium', + 'high', + 'xhigh' + ]) +}) diff --git a/src/main/acp/discovery/acp-model-catalog-probe.ts b/src/main/acp/discovery/acp-model-catalog-probe.ts index ecf6ad7ac0b..8b935339fc8 100644 --- a/src/main/acp/discovery/acp-model-catalog-probe.ts +++ b/src/main/acp/discovery/acp-model-catalog-probe.ts @@ -1,10 +1,6 @@ import { observeAcpForkSupport } from '../capabilities/acp-fork-support' -import { - assertStructuredAgentResolutionCurrent, - type StructuredAgentResolutionOperation -} from '../../native-chat/structured-agent-command-resolution' +import { assertStructuredAgentResolutionCurrent } from '../../native-chat/structured-agent-command-resolution' import { homedir } from 'node:os' -import { setTimeout as delay } from 'node:timers/promises' import { join } from 'node:path' import type { AgentCliVersionReader } from '../../agent-launch/preflight/agent-cli-version-probe' import type { AgentModelCatalogProbe } from '../../native-chat/agent-model-catalog/agent-model-catalog-store' @@ -20,40 +16,22 @@ import { acpRunFolder } from '../launch/acp-private-temp' import { acpListingDependsOnRelease, acpListingNeedsChatFolder, - acpModelListingForRelease, - type AcpSessionModelDiscovery + acpModelListingForRelease } from './acp-model-discovery-policy' import type { AcpLaunchSpec } from '../launch/acp-launch-contract' -import type { AcpSessionStarted, AcpSessionStartOptions } from '../protocol/acp-session-setup' import { checkAcpLaunchVersion } from '../launch/acp-launch-version' import { resolveAcpLaunchInvocation, type AcpLaunchInvocationDeps } from '../launch/acp-structured-launch-resolution' -import { AcpStructuredOptions } from '../options/acp-structured-options' -import type { - InitializeResponse, - SetSessionConfigOptionResponse -} from '../generated/acp-protocol.generated' -import { withTimeout } from '../../../shared/promise-timeout-fallback' -import { AcpAgentError } from '../transport/acp-errors' +import { + readSessionModels, + type AcpModelCatalogConnection, + type AcpModelCatalogSessionConnection +} from './acp-listing-session-models' +import { AcpAgentError, AcpAgentExitedError } from '../transport/acp-errors' import { AgentModelCatalogUnavailableError } from '../../native-chat/agent-model-catalog/agent-model-catalog-unavailable' - -/** Session-free capability boundary: no session methods are exposed. */ -export type AcpModelCatalogConnection = { - initialize(): Promise - requestSessionFreeExtension(method: string, params: unknown): Promise - close(): Promise -} - -export type AcpModelCatalogSessionConnection = Omit< - AcpModelCatalogConnection, - 'requestSessionFreeExtension' -> & { - start(options: AcpSessionStartOptions): Promise - setConfigOption(configId: string, value: string): Promise - deleteSession(sessionId: string): Promise -} +import { usesModelSpecificAcpEfforts } from '../../../shared/agent-session/options/agent-model-catalog-ownership' export type AcpModelCatalogProbeDeps = AcpLaunchInvocationDeps & { probeVersion?: AgentCliVersionReader @@ -113,6 +91,11 @@ export function createAcpModelCatalogProbe( envToDelete } const run = { label: spec.agent, signal: options?.signal } + // OMP records every model switch as account usage; starting on the picked model records none. + const launchedModel = + spec.launchesWithModel && usesModelSpecificAcpEfforts(spec.agent) + ? options?.selectedModels?.[0] + : undefined let models: AgentSessionModelOption[] let attemptedModels: string[] | undefined if (listing.kind === 'command') { @@ -133,32 +116,49 @@ export function createAcpModelCatalogProbe( } // Remove disposable session storage after the process closes. const folderArgs = listing.sessionFolderArgs - const folder = folderArgs - ? await acpRunFolder(join(deps.dataDirectory, 'acp-listing-sessions', spec.agent)) - : null - try { - assertStructuredAgentResolutionCurrent(operation) - const args = - folderArgs && folder ? [...launch.args, ...folderArgs(folder.path)] : launch.args - const result = await runAgentModelCatalogSession( - () => (deps.connectSession ?? createAcpAgentConnection)({ ...launch, args }), - (connection, ready) => - readSessionModels( - spec, - listing, - connection, - workspace, - operation, - options?.selectedModels, - spec.dialect.effortChangesModel ? ready : undefined - ), - run - ) - models = result.models - attemptedModels = result.attemptedModels - } finally { - await folder?.remove() + const listInSession = async (model: string | undefined) => { + const folder = folderArgs + ? await acpRunFolder(join(deps.dataDirectory, 'acp-listing-sessions', spec.agent)) + : null + try { + assertStructuredAgentResolutionCurrent(operation) + const args = [ + ...prefixArgs, + ...spec.args({ fullAccess: false, pluginDir: null, model }), + ...(folderArgs && folder ? folderArgs(folder.path) : []) + ] + return await runAgentModelCatalogSession( + () => (deps.connectSession ?? createAcpAgentConnection)({ ...launch, args }), + (connection, ready) => + readSessionModels( + spec, + listing, + connection, + workspace, + operation, + options?.selectedModels, + model, + spec.dialect.effortChangesModel ? ready : undefined + ), + run + ) + } finally { + await folder?.remove() + } } + let result: Awaited> + try { + result = await listInSession(launchedModel) + } catch (error) { + // OMP exits before ACP on a model it cannot start (a stale `:level` id); list without it. + if (launchedModel === undefined || !(error instanceof AcpAgentExitedError)) { + throw error + } + assertStructuredAgentResolutionCurrent(operation) + result = { ...(await listInSession(undefined)), attemptedModels: [launchedModel] } + } + models = result.models + attemptedModels = result.attemptedModels } else { models = await runAgentModelCatalogSession( () => (deps.connect ?? createAcpAgentConnection)(launch), @@ -190,107 +190,3 @@ export function createAcpModelCatalogProbe( const folderlessListing = (spec: AcpLaunchSpec) => new AgentModelCatalogListingStoppedError(`${spec.agent} lists in a chat's folder`) - -// OpenCode can answer -32603 while its providers are still opening; a later attempt lists them. -async function startListingSession( - connection: AcpModelCatalogSessionConnection, - workspace: string, - operation?: StructuredAgentResolutionOperation -): Promise { - for (const backoffMs of [250, 500]) { - try { - return await connection.start({ cwd: workspace, mcpServers: [] }) - } catch (error) { - if (!(error instanceof AcpAgentError && error.code === -32603)) { - throw error - } - } - await delay(backoffMs, undefined, { signal: operation?.signal }) - assertStructuredAgentResolutionCurrent(operation) - } - return connection.start({ cwd: workspace, mcpServers: [] }) -} - -async function readSessionModels( - spec: AcpLaunchSpec, - listing: AcpSessionModelDiscovery, - connection: AcpModelCatalogSessionConnection, - workspace: string, - operation?: StructuredAgentResolutionOperation, - selectedModels?: readonly string[], - ready?: (cleanup: () => Promise) => void -): Promise<{ models: AgentSessionModelOption[]; attemptedModels: string[] }> { - const initialized = await connection.initialize() - assertStructuredAgentResolutionCurrent(operation) - observeAcpForkSupport(spec.agent, initialized) - const deletes = Boolean(initialized.agentCapabilities?.sessionCapabilities?.delete) - if (!deletes && !listing.sessionFolderArgs) { - throw new Error(`${spec.agent} cannot delete a session, so it lists its models only in chats`) - } - const started = await startListingSession(connection, workspace, operation) - let stopped = false - const cleanup = async (): Promise => { - stopped = true - if (deletes) { - await connection.deleteSession(started.sessionId).catch((error: unknown) => { - console.warn(`[acp-model-catalog] ${spec.agent} kept its listing session:`, error) - }) - } - } - // Startup and menu discovery have separate budgets; cleanup also runs on cancellation. - ready?.(cleanup) - const options = new AcpStructuredOptions(spec.dialect) - options.adoptSession(started.response, 'new') - let models = options.read().models - const attemptedModels: string[] = [] - try { - if (spec.dialect.effortChangesModel) { - const byId = new Map(models.map((model) => [model.id, model])) - const targets = selectedModels ? selectedModels.flatMap((id) => byId.get(id) ?? []) : models - const sweep = async () => { - for (const model of targets) { - if (stopped) { - return - } - assertStructuredAgentResolutionCurrent(operation) - if (model.effortsKnown !== false) { - continue - } - attemptedModels.push(model.id) - try { - for (let attempt = 0; attempt < 2 && !stopped; attempt++) { - const write = options.write('model', model.id) - if (write?.method !== 'config') { - break - } - const reply = await connection.setConfigOption(write.configId, write.value) - if (stopped) { - return - } - assertStructuredAgentResolutionCurrent(operation) - options.adoptConfigOptions(reply.configOptions) - const observed = options.read() - if (observed.current.model !== model.id) { - continue - } - const known = observed.models.find((row) => row.id === model.id) - if (known) { - models = models.map((row) => (row.id === model.id ? known : row)) - } - break - } - } catch { - assertStructuredAgentResolutionCurrent(operation) - } - } - } - await withTimeout(sweep(), 5_000, undefined) - } - assertStructuredAgentResolutionCurrent(operation) - return { models, attemptedModels } - } finally { - if (!ready) { - await cleanup() - } - } -} diff --git a/src/main/acp/fixtures/omp-v18.17.0-model-efforts.json b/src/main/acp/fixtures/omp-v18.17.0-model-efforts.json new file mode 100644 index 00000000000..4ba4f7835a0 --- /dev/null +++ b/src/main/acp/fixtures/omp-v18.17.0-model-efforts.json @@ -0,0 +1,102 @@ +{ + "new": { + "configOptions": [ + { + "id": "model", + "name": "Model", + "category": "model", + "type": "select", + "currentValue": "fake/wide", + "options": [ + { "value": "fake/wide", "name": "Wide", "description": "fake/wide" }, + { "value": "fake/narrow", "name": "Narrow", "description": "fake/narrow" }, + { + "value": "apple/on-device", + "name": "Apple AFM 3 Core Advanced", + "description": "apple/on-device" + } + ] + }, + { + "id": "thinking", + "name": "Thinking", + "category": "thought_level", + "type": "select", + "currentValue": "high", + "options": [ + { "value": "off", "name": "Off" }, + { "value": "auto", "name": "Auto", "description": "Auto-detect per prompt" }, + { "value": "minimal", "name": "minimal" }, + { "value": "low", "name": "low" }, + { "value": "medium", "name": "medium" }, + { "value": "high", "name": "high" }, + { "value": "xhigh", "name": "xhigh" } + ] + } + ] + }, + "narrow": { + "configOptions": [ + { + "id": "model", + "name": "Model", + "category": "model", + "type": "select", + "currentValue": "fake/narrow", + "options": [ + { "value": "fake/wide", "name": "Wide", "description": "fake/wide" }, + { "value": "fake/narrow", "name": "Narrow", "description": "fake/narrow" }, + { + "value": "apple/on-device", + "name": "Apple AFM 3 Core Advanced", + "description": "apple/on-device" + } + ] + }, + { + "id": "thinking", + "name": "Thinking", + "category": "thought_level", + "type": "select", + "currentValue": "high", + "options": [ + { "value": "off", "name": "Off" }, + { "value": "auto", "name": "Auto", "description": "Auto-detect per prompt" }, + { "value": "low", "name": "low" }, + { "value": "high", "name": "high" } + ] + } + ] + }, + "unresolved": { + "configOptions": [ + { + "id": "model", + "name": "Model", + "category": "model", + "type": "select", + "currentValue": "fake/wide", + "options": [ + { "value": "fake/wide", "name": "Wide", "description": "fake/wide" }, + { "value": "fake/narrow", "name": "Narrow", "description": "fake/narrow" }, + { + "value": "apple/on-device", + "name": "Apple AFM 3 Core Advanced", + "description": "apple/on-device" + } + ] + }, + { + "id": "thinking", + "name": "Thinking", + "category": "thought_level", + "type": "select", + "currentValue": "off", + "options": [ + { "value": "off", "name": "Off" }, + { "value": "auto", "name": "Auto", "description": "Auto-detect per prompt" } + ] + } + ] + } +} diff --git a/src/main/acp/options/acp-structured-options.ts b/src/main/acp/options/acp-structured-options.ts index 5327bde5116..8ba7d511888 100644 --- a/src/main/acp/options/acp-structured-options.ts +++ b/src/main/acp/options/acp-structured-options.ts @@ -250,10 +250,7 @@ export class AcpStructuredOptions { model.id === currentModel ? { efforts: sessionEfforts } : (modelOption || effortOption) && !advertised - ? { - efforts: this.dialect.clampsUnlistedEffort ? sessionEfforts : [], - effortsKnown: false as const - } + ? { efforts: [], effortsKnown: false as const } : { efforts: [] } return { ...model, diff --git a/src/main/native-chat/agent-model-catalog/agent-model-catalog-effort-target-progress.test.ts b/src/main/native-chat/agent-model-catalog/agent-model-catalog-effort-target-progress.test.ts new file mode 100644 index 00000000000..01f17e28147 --- /dev/null +++ b/src/main/native-chat/agent-model-catalog/agent-model-catalog-effort-target-progress.test.ts @@ -0,0 +1,50 @@ +import { expect, it, onTestFinished, vi } from 'vitest' +import { createAgentModelCatalogService } from './agent-model-catalog-service' +import { AgentModelCatalogStore, type AgentModelCatalogSuccess } from './agent-model-catalog-store' + +it("reports a picked OMP model's menu discovery as a listing in progress", async () => { + const listings: PromiseWithResolvers[] = [] + const probe = vi.fn((_home: unknown, _options?: { selectedModels?: readonly string[] }) => { + const listing = Promise.withResolvers() + listings.push(listing) + return listing.promise + }) + const service = createAgentModelCatalogService({ + store: new AgentModelCatalogStore(), + getConversation: () => undefined, + getRecord: () => undefined, + drivesRecord: () => true, + resolveAccountHome: async () => ({ variable: 'PI_CODING_AGENT_DIR', path: '/account' }), + probes: { omp: probe } + }) + onTestFinished(() => { + for (const listing of listings) { + listing.resolve({ origin: 'probe', models: [] }) + } + return service.stop() + }) + const params = { agent: 'omp', workspacePath: '/work' } + void service.demand(params) + await vi.waitFor(() => expect(probe).toHaveBeenCalledOnce()) + listings[0]!.resolve({ + origin: 'probe', + models: [ + { id: 'wide', label: 'Wide', isDefault: false, efforts: [{ value: 'high', label: 'high' }] }, + { id: 'narrow', label: 'Narrow', isDefault: false, efforts: [], effortsKnown: false } + ] + }) + await vi.waitFor(async () => + expect((await service.read(params)).listingInProgress).toBeUndefined() + ) + + // The picker reads once, as the RPC does, and waits for the listing only if told one runs. + const picked = { ...params, selectedModel: 'narrow' } + void service.demand(picked) + expect(await service.read(picked)).toMatchObject({ listingInProgress: true }) + await vi.waitFor(() => + expect(probe).toHaveBeenLastCalledWith( + expect.anything(), + expect.objectContaining({ selectedModels: ['narrow'] }) + ) + ) +}) diff --git a/src/main/native-chat/agent-model-catalog/agent-model-catalog-entry.test.ts b/src/main/native-chat/agent-model-catalog/agent-model-catalog-entry.test.ts index a7f586a5c91..9286e5f368a 100644 --- a/src/main/native-chat/agent-model-catalog/agent-model-catalog-entry.test.ts +++ b/src/main/native-chat/agent-model-catalog/agent-model-catalog-entry.test.ts @@ -74,14 +74,16 @@ describe('a model whose newer chat listing does not know its efforts', () => { expect(merged).not.toHaveProperty('defaultEffort') }) - it('keeps native suggestions unknown without a known model menu', () => { + it('drops an OMP menu an older build borrowed from the running model', () => { + const saved = unknownChat(1, menu('low', 'medium', 'max')) + expect(agentModelCatalogEntry('omp', 'fp', saved, null, null)?.models[0]).toMatchObject({ + efforts: [], + effortsKnown: false + }) + }) + + it('keeps an older unknown menu when the newer listing has none', () => { const older = unknownChat(1, menu('low', 'high')) - const newer = unknownChat(2, menu('off', 'auto')) - for (const discovered of [null, { ...older, origin: 'probe' as const }]) { - const merged = agentModelCatalogEntry('omp', 'fp', discovered, newer, null)?.models[0] - expect(merged).toMatchObject({ efforts: menu('off', 'auto'), effortsKnown: false }) - expect(merged).not.toHaveProperty('defaultEffort') - } expect(model({ ...older, origin: 'probe' }, unknownChat(2))).toMatchObject({ efforts: menu('low', 'high'), effortsKnown: false diff --git a/src/main/native-chat/agent-model-catalog/agent-model-catalog-models.ts b/src/main/native-chat/agent-model-catalog/agent-model-catalog-models.ts index d9945ac0782..3b6ffd9e712 100644 --- a/src/main/native-chat/agent-model-catalog/agent-model-catalog-models.ts +++ b/src/main/native-chat/agent-model-catalog/agent-model-catalog-models.ts @@ -28,7 +28,7 @@ export function mergedAgentModelCatalogModels( // Coarser probe menus must not discard a chat-confirmed effort unless a newer chat menu does. const keepsChatMenu = older === live && !offersConfigured(model.efforts) && offersConfigured(olderEfforts) - const { efforts, effortsKnown } = modelSpecificEfforts + const { efforts: keptEfforts, effortsKnown } = modelSpecificEfforts ? model.effortsKnown === false && olderModel ? olderModel : model @@ -39,6 +39,8 @@ export function mergedAgentModelCatalogModels( : model.efforts.length > 0 && !keepsChatMenu ? model : (olderModel ?? model) + // An unobserved per-model menu is another model's; older builds saved such copies. + const efforts = modelSpecificEfforts && effortsKnown === false ? [] : keptEfforts const defaultEffort = [configuredEffort, listed?.defaultEffort, reported?.defaultEffort].find( (effort) => effort !== undefined && efforts.some((choice) => choice.value === effort) ) diff --git a/src/main/native-chat/agent-model-catalog/agent-model-catalog-snapshot.ts b/src/main/native-chat/agent-model-catalog/agent-model-catalog-snapshot.ts index b737ff65017..55cb348482b 100644 --- a/src/main/native-chat/agent-model-catalog/agent-model-catalog-snapshot.ts +++ b/src/main/native-chat/agent-model-catalog/agent-model-catalog-snapshot.ts @@ -239,7 +239,10 @@ export function projectAgentModelCatalogSnapshot( ...(!params.savedOnly && !params.verifiedOnly && (deps.store.isListing(bound.fingerprint) || - (resolving && deps.store.probeDue(bound.fingerprint))) && + (resolving && + (deps.store.probeDue(bound.fingerprint) || + // The picked model's menu is listed by the refresh this read just asked for. + deps.store.efforts.due(bound.fingerprint, params.selectedModel)))) && (independent || !listed || usesModelSpecificAcpEfforts(params.agent)) ? { listingInProgress: true as const } : {}), diff --git a/src/shared/agent-session/options/agent-model-catalog-ownership.ts b/src/shared/agent-session/options/agent-model-catalog-ownership.ts index 601dd3323b1..9136e84c4d3 100644 --- a/src/shared/agent-session/options/agent-model-catalog-ownership.ts +++ b/src/shared/agent-session/options/agent-model-catalog-ownership.ts @@ -3,7 +3,7 @@ export function independentAgentModelCatalog(agent: string): boolean { } export function usesModelSpecificAcpEfforts(agent: string): boolean { - return agent === 'opencode' || agent === 'opencode2' + return agent === 'opencode' || agent === 'opencode2' || agent === 'omp' } export function usesInstalledAgentModelCatalog(agent: string, coherentCatalog?: boolean): boolean { From 08bb458da43ca97c39443444ff88368cb3e90928 Mon Sep 17 00:00:00 2001 From: Neil <4138956+nwparker@users.noreply.github.com> Date: Tue, 10 Nov 2026 16:34:24 -0800 Subject: [PATCH 04/21] Remove the per-repo agent folder trust prompt (#40934) The first-launch 'Trust for agents?' modal blocked agent launches to ask a question that offered no real protection: undecided repos, and every unattended launch, already pre-trusted. Drop the dialog, the launch hold and the skipAgentWorkspaceTrust plumbing. Repos default to pre-trust; the per-repo setting remains as an opt-out. --- .../trust/agent-workspace-trust-spawn.test.ts | 8 - .../trust/agent-workspace-trust-spawn.ts | 3 - src/main/ipc/pty/ipc/spawn-types.ts | 2 - src/main/ipc/pty/spawn-options-assembly.ts | 2 - src/preload/api/pty-api.ts | 2 - src/preload/api/pty-bridge-session-control.ts | 2 - src/renderer/src/App.tsx | 2 - ...ntWorkspaceTrustDecisionDialog.fixture.tsx | 16 -- .../AgentWorkspaceTrustDecisionDialog.tsx | 106 ----------- .../agent-workspace-trust-decision-copy.ts | 21 --- ...nt-workspace-trust-decision-prompt.test.ts | 148 --------------- .../agent-workspace-trust-decision-prompt.ts | 169 ------------------ .../behavior/agent-workspace-trust-copy.ts | 2 +- .../RepositoryAgentTrustSection.fixture.tsx | 1 - .../RepositoryAgentTrustSection.tsx | 8 +- .../pty-transport/ipc/ipc-pty-connect.ts | 23 +-- .../ipc/ipc-pty-spawn-request.ts | 5 +- src/renderer/src/i18n/locales/de.json | 6 +- src/renderer/src/i18n/locales/en.json | 6 +- src/renderer/src/i18n/locales/es.json | 6 +- src/renderer/src/i18n/locales/fr.json | 6 +- src/renderer/src/i18n/locales/ja.json | 6 +- src/renderer/src/i18n/locales/ko.json | 6 +- src/renderer/src/i18n/locales/pt.json | 6 +- src/renderer/src/i18n/locales/ru.json | 6 +- src/renderer/src/i18n/locales/tr.json | 6 +- src/renderer/src/i18n/locales/zh-Hant.json | 6 +- src/renderer/src/i18n/locales/zh.json | 6 +- .../launch-parity-renderer.test-fixture.ts | 4 +- .../agent-workspace-trust-decision.ts | 12 +- .../languages/protected-copy-keys.json | 4 - ...agent-workspace-trust-first-launch.spec.ts | 84 --------- tests/e2e/helpers/orca-app.ts | 8 +- tests/e2e/helpers/orca-restart.ts | 6 +- 34 files changed, 23 insertions(+), 681 deletions(-) delete mode 100644 src/renderer/src/components/agent-workspace-trust/AgentWorkspaceTrustDecisionDialog.fixture.tsx delete mode 100644 src/renderer/src/components/agent-workspace-trust/AgentWorkspaceTrustDecisionDialog.tsx delete mode 100644 src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-prompt.test.ts delete mode 100644 src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-prompt.ts delete mode 100644 tests/e2e/agent-workspace-trust-first-launch.spec.ts diff --git a/src/main/agent-launch/trust/agent-workspace-trust-spawn.test.ts b/src/main/agent-launch/trust/agent-workspace-trust-spawn.test.ts index 5bcda0cae73..14cc1d4fec8 100644 --- a/src/main/agent-launch/trust/agent-workspace-trust-spawn.test.ts +++ b/src/main/agent-launch/trust/agent-workspace-trust-spawn.test.ts @@ -8,12 +8,10 @@ async function sshTrustRequest(input: { worktreeId: string globalEnabled?: boolean repoDecision?: AgentWorkspaceTrustDecision - skipForThisLaunch?: boolean }): Promise { const spawnOptions: AgentTrustSpawnFields = {} await applyAgentWorkspaceTrustToSpawn({ launchAgent: 'claude', - skipForThisLaunch: input.skipForThisLaunch, worktreeId: input.worktreeId, cwd: '/srv/app', store: { @@ -56,10 +54,4 @@ describe('agent folder trust follows the repo decision', () => { }) ).toEqual({}) }) - - it("honours this launch's let-agents-ask answer even when the repo has none stored", async () => { - expect( - await sshTrustRequest({ worktreeId: 'repo-1::/srv/app', skipForThisLaunch: true }) - ).toEqual({}) - }) }) diff --git a/src/main/agent-launch/trust/agent-workspace-trust-spawn.ts b/src/main/agent-launch/trust/agent-workspace-trust-spawn.ts index 106da0081d8..aa37a803eaa 100644 --- a/src/main/agent-launch/trust/agent-workspace-trust-spawn.ts +++ b/src/main/agent-launch/trust/agent-workspace-trust-spawn.ts @@ -43,8 +43,6 @@ export function agentWorkspaceTrustDecisionForWorkspace( export function applyAgentWorkspaceTrustToSpawn( args: { launchAgent: unknown - /** The user chose "let agents ask" for this launch; it wins even if saving that answer failed. */ - skipForThisLaunch?: boolean /** Placement workspace id; its root may differ from the agent’s starting folder. */ worktreeId: string | undefined /** Section root to pre-trust for this launch; never part of the terminal identity. */ @@ -64,7 +62,6 @@ export function applyAgentWorkspaceTrustToSpawn( const trustWorkspaceId = resolveAgentWorkspaceTrustOwner(args) if ( !args.isFreshLaunch || - args.skipForThisLaunch === true || !shouldPreTrustAgentWorkspace( args.settings?.agentWorkspaceTrustEnabled, agentWorkspaceTrustDecisionForWorkspace(trustWorkspaceId, args.store) diff --git a/src/main/ipc/pty/ipc/spawn-types.ts b/src/main/ipc/pty/ipc/spawn-types.ts index 1a52a464b4c..08bd61db360 100644 --- a/src/main/ipc/pty/ipc/spawn-types.ts +++ b/src/main/ipc/pty/ipc/spawn-types.ts @@ -65,8 +65,6 @@ export type PtySpawnIpcArgs = { spawnIntent?: unknown // Only `true` counts: a window restore no user act asked for. wslPassiveRestore?: unknown - // Only `true` counts: the user just chose to let agents ask, even if saving that answer failed. - skipAgentWorkspaceTrust?: unknown // Validate untyped launch input through AgentLaunch.safeParse. agentLaunch?: unknown // Main validates renderer telemetry; retain unknown input types at the IPC boundary. diff --git a/src/main/ipc/pty/spawn-options-assembly.ts b/src/main/ipc/pty/spawn-options-assembly.ts index ba8e3d5b1a0..93fcecc7e12 100644 --- a/src/main/ipc/pty/spawn-options-assembly.ts +++ b/src/main/ipc/pty/spawn-options-assembly.ts @@ -93,7 +93,6 @@ export type PtySpawnLaunchOptionsInput = { connectionId?: string | null shellOverride?: string telemetry?: { agent_kind?: unknown } - skipAgentWorkspaceTrust?: unknown } env: Record | undefined /** Mutated: OpenCode prep may add the keys its launch must drop. */ @@ -210,7 +209,6 @@ export async function assemblePtySpawnLaunchOptions(input: PtySpawnLaunchOptions launchAgent: args.launchAgent, worktreeId: args.worktreeId, trustWorktreeId: args.trustWorktreeId, - skipForThisLaunch: args.skipAgentWorkspaceTrust === true, cwd: input.cwd, store: input.store, isFreshLaunch: !input.preAdoptedStablePane && launchCommand !== undefined, diff --git a/src/preload/api/pty-api.ts b/src/preload/api/pty-api.ts index 68e22c51ffc..7dba2b56a1b 100644 --- a/src/preload/api/pty-api.ts +++ b/src/preload/api/pty-api.ts @@ -79,8 +79,6 @@ export type PtyApi = { spawnIntent?: SpawnIntent // A restore no user act asked for; main holds it for a stopped WSL distro. Older mains ignore it. wslPassiveRestore?: true - // The user's "let agents ask" answer for this launch; older mains ignore it. - skipAgentWorkspaceTrust?: true agentLaunch?: AgentPaneLaunchRequest // Main emits agent_started only after a successful spawn. telemetry?: { agent_kind: AgentKind; launch_source: LaunchSource; request_kind: RequestKind } diff --git a/src/preload/api/pty-bridge-session-control.ts b/src/preload/api/pty-bridge-session-control.ts index b57e1593f61..685259339ec 100644 --- a/src/preload/api/pty-bridge-session-control.ts +++ b/src/preload/api/pty-bridge-session-control.ts @@ -66,8 +66,6 @@ export const ptySessionControlApi = { placement?: TerminalPanePlacement spawnIntent?: SpawnIntent wslPassiveRestore?: true - // The user's "let agents ask" answer for this launch; older mains ignore it. - skipAgentWorkspaceTrust?: true agentLaunch?: AgentPaneLaunchRequest telemetry?: { agent_kind: AgentKind; launch_source: LaunchSource; request_kind: RequestKind } }): Promise<{ diff --git a/src/renderer/src/App.tsx b/src/renderer/src/App.tsx index 403dfb08d56..b4d2e3e3fec 100644 --- a/src/renderer/src/App.tsx +++ b/src/renderer/src/App.tsx @@ -11,7 +11,6 @@ import { SkillFreshnessNudge } from './components/skills/freshness/SkillFreshnes import PinnedTabCloseDialog from './components/terminal-pane/layout/close/PinnedTabCloseDialog' import RunningTerminalCloseDialog from './components/terminal-pane/layout/close/RunningTerminalCloseDialog' import WorktreeBaseFallbackDialog from './components/WorktreeBaseFallbackDialog' -import AgentWorkspaceTrustDecisionDialog from './components/agent-workspace-trust/AgentWorkspaceTrustDecisionDialog' import ClaudeAccountSwitchDialog from './components/ClaudeAccountSwitchDialog' import { useUnreadDockBadge } from './hooks/useUnreadDockBadge' import { AppBackgroundServices } from './app-shell/AppBackgroundServices' @@ -128,7 +127,6 @@ function App(): React.JSX.Element { - diff --git a/src/renderer/src/components/agent-workspace-trust/AgentWorkspaceTrustDecisionDialog.fixture.tsx b/src/renderer/src/components/agent-workspace-trust/AgentWorkspaceTrustDecisionDialog.fixture.tsx deleted file mode 100644 index 5394cf571e6..00000000000 --- a/src/renderer/src/components/agent-workspace-trust/AgentWorkspaceTrustDecisionDialog.fixture.tsx +++ /dev/null @@ -1,16 +0,0 @@ -import { Dialog } from '@/components/ui/dialog' -import { defineFixture } from '../../component-fixtures/component-fixture' -import { AgentWorkspaceTrustDecisionDialogView } from './AgentWorkspaceTrustDecisionDialog' - -export const firstLaunch = defineFixture({ - width: 640, - render: () => ( - - {}} - /> - - ) -}) diff --git a/src/renderer/src/components/agent-workspace-trust/AgentWorkspaceTrustDecisionDialog.tsx b/src/renderer/src/components/agent-workspace-trust/AgentWorkspaceTrustDecisionDialog.tsx deleted file mode 100644 index 1fce90c9492..00000000000 --- a/src/renderer/src/components/agent-workspace-trust/AgentWorkspaceTrustDecisionDialog.tsx +++ /dev/null @@ -1,106 +0,0 @@ -import { useEffect, useState, useSyncExternalStore } from 'react' -import { ShieldQuestion } from 'lucide-react' -import { Button } from '@/components/ui/button' -import { useAppStore } from '@/store' -import { - Dialog, - DialogContent, - DialogDescription, - DialogFooter, - DialogHeader, - DialogTitle -} from '@/components/ui/dialog' -import type { AgentWorkspaceTrustDecision } from '../../../../shared/agent-launch/permissions/agent-workspace-trust-decision' -import { agentWorkspaceTrustDecisionCopy as copy } from './agent-workspace-trust-decision-copy' -import { - answerAgentWorkspaceTrustDecision, - getAgentWorkspaceTrustDecisionRequest, - subscribeAgentWorkspaceTrustDecisionRequest -} from './agent-workspace-trust-decision-prompt' - -export type AgentWorkspaceTrustDecisionDialogViewProps = { - repoName: string - saving: boolean - onAnswer: (decision: AgentWorkspaceTrustDecision) => void -} - -/** A launch waits on this answer, so it closes only through a choice. */ -export function AgentWorkspaceTrustDecisionDialogView({ - repoName, - saving, - onAnswer -}: AgentWorkspaceTrustDecisionDialogViewProps): React.JSX.Element { - return ( - event.preventDefault()} - onInteractOutside={(event) => event.preventDefault()} - > - - - - - - {copy.promptDescription()} - -

{copy.promptRemembered()}

- - - - -
- ) -} - -export default function AgentWorkspaceTrustDecisionDialog(): React.JSX.Element | null { - const request = useSyncExternalStore( - subscribeAgentWorkspaceTrustDecisionRequest, - getAgentWorkspaceTrustDecisionRequest, - getAgentWorkspaceTrustDecisionRequest - ) - const [savingFor, setSavingFor] = useState(null) - const setContextualToursBlockingSurfaceVisible = useAppStore( - (state) => state.setContextualToursBlockingSurfaceVisible - ) - const open = request !== null - useEffect(() => { - setContextualToursBlockingSurfaceVisible(open) - return () => setContextualToursBlockingSurfaceVisible(false) - }, [open, setContextualToursBlockingSurfaceVisible]) - if (!request) { - return null - } - const answer = (decision: AgentWorkspaceTrustDecision): void => { - setSavingFor(request) - void answerAgentWorkspaceTrustDecision(decision).finally(() => setSavingFor(null)) - } - return ( - - - - ) -} diff --git a/src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-copy.ts b/src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-copy.ts index 8306ca00ed1..70aadfde427 100644 --- a/src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-copy.ts +++ b/src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-copy.ts @@ -2,22 +2,6 @@ import { translate } from '@/i18n/i18n' export const agentWorkspaceTrustDecisionCopy = { - promptTitle: (repo: string) => - translate( - 'auto.components.agent-workspace-trust-decision.prompt-title', - 'Trust {{repo}} for agents?', - { repo } - ), - promptDescription: () => - translate( - 'auto.components.agent-workspace-trust-decision.prompt-description', - 'Agents like Claude Code and Codex ask "Do you trust this folder?" before they load the project\'s own agent settings and hooks, which can run commands. Orca can answer yes for you in every worktree of this repository.' - ), - promptRemembered: () => - translate( - 'auto.components.agent-workspace-trust-decision.prompt-remembered', - "Orca remembers your answer for this repository. You can change it in the repository's settings." - ), trustRepository: () => translate( 'auto.components.agent-workspace-trust-decision.trust-repository', @@ -32,11 +16,6 @@ export const agentWorkspaceTrustDecisionCopy = { 'auto.components.agent-workspace-trust-decision.setting-description', 'Whether agents Orca starts in this repository\'s worktrees skip their "Do you trust this folder?" prompt and load the project\'s agent settings and hooks right away.' ), - settingUndecided: () => - translate( - 'auto.components.agent-workspace-trust-decision.setting-undecided', - 'Not decided yet. Orca will ask the next time you start an agent here.' - ), settingGlobalOff: () => translate( 'auto.components.agent-workspace-trust-decision.setting-global-off', diff --git a/src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-prompt.test.ts b/src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-prompt.test.ts deleted file mode 100644 index 8d1358d857f..00000000000 --- a/src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-prompt.test.ts +++ /dev/null @@ -1,148 +0,0 @@ -import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest' -import type { Repo } from '../../../../shared/repo-types' -import { - installIpcPtyWindow, - restorePtySpecWindow -} from '../terminal-pane/pty-transport/ipc/pty-transport-test-harness' - -type StoreState = { - repos: Partial[] - settings: { agentWorkspaceTrustEnabled: boolean } - updateRepo: ReturnType -} - -const state = vi.hoisted((): StoreState => ({ - repos: [], - settings: { agentWorkspaceTrustEnabled: true }, - updateRepo: vi.fn() -})) - -vi.mock('@/store', () => ({ useAppStore: { getState: () => state } })) - -import { - AGENT_TRUST_PROMPT_TIMEOUT_MS, - AGENT_TRUST_SAVE_WAIT_MS, - answerAgentWorkspaceTrustDecision, - awaitAgentWorkspaceTrustDecision, - getAgentWorkspaceTrustDecisionRequest, - subscribeAgentWorkspaceTrustDecisionRequest -} from './agent-workspace-trust-decision-prompt' - -const WORKTREE = 'repo-1::/repos/app/feature' -const launch = (workspaceId = WORKTREE) => - awaitAgentWorkspaceTrustDecision({ launchAgent: 'claude', workspaceId, connectionId: null }) - -let unsubscribeDialog: () => void = () => {} - -beforeEach(() => { - vi.useFakeTimers() - vi.spyOn(console, 'warn').mockImplementation(() => {}) - unsubscribeDialog = subscribeAgentWorkspaceTrustDecisionRequest(() => {}) - state.repos = [{ id: 'repo-1', path: '/repos/app', displayName: 'app' }] - state.settings = { agentWorkspaceTrustEnabled: true } - state.updateRepo = vi.fn(async (repoId: string, updates: Partial) => { - state.repos = state.repos.map((repo) => (repo.id === repoId ? { ...repo, ...updates } : repo)) - return true - }) -}) - -afterEach(async () => { - // Release anything a test left on screen so module state never leaks between tests. - await vi.advanceTimersByTimeAsync(AGENT_TRUST_PROMPT_TIMEOUT_MS) - unsubscribeDialog() - vi.useRealTimers() - vi.restoreAllMocks() -}) - -describe('first agent launch in a repo', () => { - it.each([true, false])( - 'asks once for concurrent section launches through PTY connect (save succeeds: %s)', - async (saveSucceeds) => { - const originalWindow: typeof window | undefined = globalThis.window - installIpcPtyWindow(originalWindow, {}) - if (!saveSucceeds) { - state.updateRepo = vi.fn(async () => false) - } - vi.mocked(window.api.pty.spawn) - .mockResolvedValueOnce({ id: 'section-1' }) - .mockResolvedValueOnce({ id: 'section-2' }) - const { createIpcPtyTransport } = - await import('../terminal-pane/pty-transport/ipc/pty-transport') - const transports = ['folder:placement', 'folder:other'].map((worktreeId) => - createIpcPtyTransport({ - worktreeId, - trustWorktreeId: WORKTREE, - cwd: '/repos/app/feature', - command: 'claude', - launchAgent: 'claude' - }) - ) - const connecting = Promise.all( - transports.map((transport) => transport.connect({ url: '', callbacks: {} })) - ) - try { - expect(getAgentWorkspaceTrustDecisionRequest()).toMatchObject({ - repoId: 'repo-1', - repoName: 'app', - hostId: 'local' - }) - expect(window.api.pty.spawn).not.toHaveBeenCalled() - await answerAgentWorkspaceTrustDecision('agent-asks') - await connecting - expect(state.updateRepo).toHaveBeenCalledExactlyOnceWith( - 'repo-1', - { agentWorkspaceTrust: 'agent-asks' }, - { hostId: 'local' } - ) - expect(getAgentWorkspaceTrustDecisionRequest()).toBeNull() - expect(vi.mocked(window.api.pty.spawn).mock.calls.map(([args]) => args)).toMatchObject([ - { - worktreeId: 'folder:placement', - trustWorktreeId: WORKTREE, - cwd: '/repos/app/feature', - skipAgentWorkspaceTrust: true - }, - { - worktreeId: 'folder:other', - trustWorktreeId: WORKTREE, - cwd: '/repos/app/feature', - skipAgentWorkspaceTrust: true - } - ]) - } finally { - await answerAgentWorkspaceTrustDecision('agent-asks') - await connecting - for (const transport of transports) { - transport.destroy?.() - } - restorePtySpecWindow(originalWindow) - } - } - ) - - it('releases with the answer when saving stalls', async () => { - state.updateRepo = vi.fn(() => new Promise(() => {})) - const pending = launch() - - const answered = answerAgentWorkspaceTrustDecision('agent-asks') - await vi.advanceTimersByTimeAsync(AGENT_TRUST_SAVE_WAIT_MS) - await answered - - await expect(pending).resolves.toBe('agent-asks') - }) - - it("releases an unanswered prompt with today's default and closes it", async () => { - const pending = launch() - - await vi.advanceTimersByTimeAsync(AGENT_TRUST_PROMPT_TIMEOUT_MS) - - await expect(pending).resolves.toBeNull() - expect(getAgentWorkspaceTrustDecisionRequest()).toBeNull() - expect(state.updateRepo).not.toHaveBeenCalled() - }) - - it('never holds a launch when no dialog is mounted to answer it', () => { - unsubscribeDialog() - expect(launch()).toBeNull() - }) -}) diff --git a/src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-prompt.ts b/src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-prompt.ts deleted file mode 100644 index ee8bd968bac..00000000000 --- a/src/renderer/src/components/agent-workspace-trust/agent-workspace-trust-decision-prompt.ts +++ /dev/null @@ -1,169 +0,0 @@ -import { useAppStore } from '@/store' -import { getRepoRowIdentity, repoMatchesHostIdentity } from '@/store/repos/repo-host-identity' -import { - TUI_AGENT_CONFIG, - isTuiAgent -} from '../../../../shared/agent-launch/catalog/tui-agent-config' -import { - needsAgentWorkspaceTrustDecision, - type AgentWorkspaceTrustDecision -} from '../../../../shared/agent-launch/permissions/agent-workspace-trust-decision' -import { - getConnectionExecutionHostId, - getRepoExecutionHostId -} from '../../../../shared/execution-host' -import type { Repo } from '../../../../shared/repo-types' -import { parseWorkspaceKey } from '../../../../shared/workspace-scope' -import { splitWorktreeId } from '../../../../shared/worktree/id' - -export type AgentWorkspaceTrustDecisionRequest = { - repoId: string - repoName: string - hostId: ReturnType -} - -/** The user's answer for this launch; null means no answer, so today's default applies. */ -export type AgentWorkspaceTrustLaunchAnswer = AgentWorkspaceTrustDecision | null - -/** An unanswered prompt releases its launches with today's default rather than holding them. */ -export const AGENT_TRUST_PROMPT_TIMEOUT_MS = 120_000 -/** A slow save must not hold the launch; the answer still applies to this launch. */ -export const AGENT_TRUST_SAVE_WAIT_MS = 5_000 - -type PendingRequest = AgentWorkspaceTrustDecisionRequest & { - rowIdentity: string - settle: (answer: AgentWorkspaceTrustLaunchAnswer) => void - timer?: ReturnType - answering?: boolean -} - -const queue: PendingRequest[] = [] -const inFlightByRow = new Map>() -const listeners = new Set<() => void>() - -function notifyListeners(): void { - for (const listener of listeners) { - listener() - } -} - -function release(request: PendingRequest, answer: AgentWorkspaceTrustLaunchAnswer): void { - clearTimeout(request.timer) - const index = queue.indexOf(request) - if (index !== -1) { - queue.splice(index, 1) - } - inFlightByRow.delete(request.rowIdentity) - request.settle(answer) - showNext() -} - -/** The timeout starts when a request is on screen, so a queued repo is not timed out unseen. */ -function showNext(): void { - const next = queue[0] - if (next && next.timer === undefined) { - next.timer = setTimeout(() => { - if (!next.answering) { - release(next, null) - } - }, AGENT_TRUST_PROMPT_TIMEOUT_MS) - } - notifyListeners() -} - -/** The PTY's own connection names the owning host; never fall back to the focused host. */ -function repoForWorkspace( - workspaceId: string | undefined, - connectionId: string | null | undefined -): Repo | null { - const scope = workspaceId ? parseWorkspaceKey(workspaceId) : null - if (!workspaceId || scope?.type === 'folder') { - return null - } - const repoId = splitWorktreeId( - scope?.type === 'worktree' ? scope.worktreeId : workspaceId - )?.repoId - if (!repoId) { - return null - } - const hostId = getConnectionExecutionHostId(connectionId) - return ( - useAppStore.getState().repos.find((repo) => repoMatchesHostIdentity(repo, repoId, hostId)) ?? - null - ) -} - -/** Null when no answer is needed, so launches that need none keep their synchronous path. */ -export function awaitAgentWorkspaceTrustDecision(args: { - launchAgent: unknown - workspaceId: string | undefined - connectionId: string | null | undefined -}): Promise | null { - // No mounted dialog could answer, so never hold the launch. - if (listeners.size === 0) { - return null - } - if (!isTuiAgent(args.launchAgent) || !TUI_AGENT_CONFIG[args.launchAgent].preflightTrust) { - return null - } - const repo = repoForWorkspace(args.workspaceId, args.connectionId) - const globalEnabled = useAppStore.getState().settings?.agentWorkspaceTrustEnabled - if (!repo || !needsAgentWorkspaceTrustDecision(globalEnabled, repo.agentWorkspaceTrust)) { - return null - } - const rowIdentity = getRepoRowIdentity(repo) - const existing = inFlightByRow.get(rowIdentity) - if (existing) { - return existing - } - const pending = new Promise((settle) => { - queue.push({ - repoId: repo.id, - repoName: repo.displayName, - hostId: getRepoExecutionHostId(repo), - rowIdentity, - settle - }) - }) - inFlightByRow.set(rowIdentity, pending) - showNext() - return pending -} - -/** Saves the answer, then releases waiting launches with it even if the save fails or stalls. */ -export async function answerAgentWorkspaceTrustDecision( - decision: AgentWorkspaceTrustDecision -): Promise { - const request = queue[0] - if (!request || request.answering) { - return - } - request.answering = true - clearTimeout(request.timer) - let saveWait: ReturnType | undefined - const saved = await Promise.race([ - useAppStore - .getState() - .updateRepo(request.repoId, { agentWorkspaceTrust: decision }, { hostId: request.hostId }) - .catch(() => false), - new Promise((resolve) => { - saveWait = setTimeout(() => resolve(false), AGENT_TRUST_SAVE_WAIT_MS) - }) - ]) - clearTimeout(saveWait) - if (!saved) { - console.warn( - '[agent-trust] the repository answer was not saved; it applies to this launch only' - ) - } - release(request, decision) -} - -export function getAgentWorkspaceTrustDecisionRequest(): AgentWorkspaceTrustDecisionRequest | null { - return queue[0] ?? null -} - -export function subscribeAgentWorkspaceTrustDecisionRequest(listener: () => void): () => void { - listeners.add(listener) - return () => listeners.delete(listener) -} diff --git a/src/renderer/src/components/settings/agents/behavior/agent-workspace-trust-copy.ts b/src/renderer/src/components/settings/agents/behavior/agent-workspace-trust-copy.ts index ad0ec43c639..a82be27c348 100644 --- a/src/renderer/src/components/settings/agents/behavior/agent-workspace-trust-copy.ts +++ b/src/renderer/src/components/settings/agents/behavior/agent-workspace-trust-copy.ts @@ -13,7 +13,7 @@ export function getAgentWorkspaceTrustTitle(t: Translate = translate): string { export function getAgentWorkspaceTrustDescription(t: Translate = translate): string { return t( 'auto.components.settings.agent-workspace-trust-copy.description', - 'Agents Orca starts can skip their "Do you trust this folder?" prompt, so the project\'s agent hooks and settings run right away. Orca asks once per repository the first time you start an agent there, and you can change the answer in that repository\'s settings. Turn this off to let agents always ask. An agent that asks while running without you watching, such as an orchestration worker, an automation or an agent started from your phone, waits until you answer in its terminal. Folders already trusted stay trusted.' + 'Agents Orca starts can skip their "Do you trust this folder?" prompt, so the project\'s agent hooks and settings run right away. You can turn this off for one repository in that repository\'s settings. Turn this off to let agents always ask. An agent that asks while running without you watching, such as an orchestration worker, an automation or an agent started from your phone, waits until you answer in its terminal. Folders already trusted stay trusted.' ) } diff --git a/src/renderer/src/components/settings/repository/RepositoryAgentTrustSection.fixture.tsx b/src/renderer/src/components/settings/repository/RepositoryAgentTrustSection.fixture.tsx index a66f2bb437b..8396190cc78 100644 --- a/src/renderer/src/components/settings/repository/RepositoryAgentTrustSection.fixture.tsx +++ b/src/renderer/src/components/settings/repository/RepositoryAgentTrustSection.fixture.tsx @@ -28,7 +28,6 @@ function fixture(decision: AgentWorkspaceTrustDecision | undefined, globalEnable }) } -export const undecided = fixture(undefined) export const preTrust = fixture('pre-trust') export const agentAsks = fixture('agent-asks') export const globalOff = fixture('pre-trust', false) diff --git a/src/renderer/src/components/settings/repository/RepositoryAgentTrustSection.tsx b/src/renderer/src/components/settings/repository/RepositoryAgentTrustSection.tsx index 279cd432107..480ae0404af 100644 --- a/src/renderer/src/components/settings/repository/RepositoryAgentTrustSection.tsx +++ b/src/renderer/src/components/settings/repository/RepositoryAgentTrustSection.tsx @@ -30,7 +30,7 @@ export function RepositoryAgentTrustSection({ (state) => state.settings?.agentWorkspaceTrustEnabled !== false ) const entry = getRepositoryAgentTrustSearchEntry() - const value = repo.agentWorkspaceTrust ?? 'undecided' + const value = repo.agentWorkspaceTrust ?? 'pre-trust' return ( {entry.title}

{entry.description}

- + value={value} onChange={(next) => { - if (next !== 'undecided' && next !== value) { + if (next !== value) { void updateRepo(repo.id, { agentWorkspaceTrust: next }) } }} @@ -59,8 +59,6 @@ export function RepositoryAgentTrustSection({ /> {!globalTrustEnabled ? (

{copy.settingGlobalOff()}

- ) : value === 'undecided' ? ( -

{copy.settingUndecided()}

) : null}
) diff --git a/src/renderer/src/components/terminal-pane/pty-transport/ipc/ipc-pty-connect.ts b/src/renderer/src/components/terminal-pane/pty-transport/ipc/ipc-pty-connect.ts index 3b92006c938..126a05e4e64 100644 --- a/src/renderer/src/components/terminal-pane/pty-transport/ipc/ipc-pty-connect.ts +++ b/src/renderer/src/components/terminal-pane/pty-transport/ipc/ipc-pty-connect.ts @@ -1,5 +1,4 @@ import { admitTerminalNoticeLifetime } from '../../provider-notices/terminal-notice-lifetime' -import { resolveAgentWorkspaceTrustOwner } from '../../../../../../shared/agent-launch/permissions/agent-workspace-trust-decision' import { terminalLaunchFacts } from '../../../../../../shared/agent-session/lifecycle/agent-session-resume' import { readTerminalSpawnDeferral } from '../../../../../../shared/terminal/pty/admission/spawn-admission-contract' import { isRuntimeOwnedSshTargetId } from '../../../../../../shared/execution-host' @@ -20,7 +19,6 @@ import type { IpcPtySessionHandlers } from './ipc-pty-session-handlers' import { isSshSessionGoneError } from '../host-ownership/ssh-session-gone-verdict' import { spawnIpcPty } from './ipc-pty-spawn-request' import { settleAgentLaunchPaneStart } from '@/lib/agent-session/launch/pane/agent-launch-pane-start' -import { awaitAgentWorkspaceTrustDecision } from '../../../agent-workspace-trust/agent-workspace-trust-decision-prompt' import type { IpcPtyTransportOptions, PtyConnectResult, PtyTransport } from '../pty-transport-types' const SSH_PTY_CONNECTION_MISMATCH_MARKER = 'belongs to SSH connection' @@ -83,31 +81,12 @@ export async function connectIpcPty( return } } - // A fresh agent launch in an undecided repo waits for the user's trust answer. - const trustDecision = - admittedSessionId || (options.command ?? transportOptions.command) === undefined - ? null - : awaitAgentWorkspaceTrustDecision({ - launchAgent: options.launchAgent ?? transportOptions.launchAgent, - workspaceId: resolveAgentWorkspaceTrustOwner(transportOptions), - connectionId: transportOptions.connectionId - }) - // Bounded: an unanswered prompt releases with today's default (agent-workspace-trust-decision-prompt). - const trustAnswer = trustDecision ? await trustDecision : null - if (trustDecision && context.isDestroyed()) { - return - } if (options.shouldContinue && !options.shouldContinue()) { return } // Fence before spawning: redeployed relays recycle IDs, so earlier buffered state belongs to a previous PTY lifetime. const priorIncarnationFence = currentPreHandlerPtySequence() - const spawnResult = await spawnIpcPty( - transportOptions, - options, - admittedSessionId, - trustAnswer === 'agent-asks' - ) + const spawnResult = await spawnIpcPty(transportOptions, options, admittedSessionId) settleAgentLaunchPaneStart(transportOptions.tabId, transportOptions.leafId, spawnResult.id) // Any resolved spawn or cold restore is live under this id, even if this transport was destroyed meanwhile. markTerminalProviderSessionBound(spawnResult.id) diff --git a/src/renderer/src/components/terminal-pane/pty-transport/ipc/ipc-pty-spawn-request.ts b/src/renderer/src/components/terminal-pane/pty-transport/ipc/ipc-pty-spawn-request.ts index 3b391c49724..76c069488c7 100644 --- a/src/renderer/src/components/terminal-pane/pty-transport/ipc/ipc-pty-spawn-request.ts +++ b/src/renderer/src/components/terminal-pane/pty-transport/ipc/ipc-pty-spawn-request.ts @@ -19,9 +19,7 @@ export type IpcPtySpawnResponse = PtyConnectResult & { export async function spawnIpcPty( transportOptions: IpcPtyTransportOptions, connectOptions: PtyConnectOptions, - admittedSessionId?: string, - /** The user's "let agents ask" answer for this launch. */ - skipAgentWorkspaceTrust = false + admittedSessionId?: string ): Promise { if ( !admittedSessionId && @@ -146,7 +144,6 @@ export async function spawnIpcPty( ...(placement && !admittedSessionId ? { placement } : {}), ...(spawnIntent ? { spawnIntent } : {}), ...(connectOptions.wslPassiveRestore ? { wslPassiveRestore: true as const } : {}), - ...(skipAgentWorkspaceTrust ? { skipAgentWorkspaceTrust: true as const } : {}), ...(replacesPtyId ? { replacesPtyId } : {}), ...(shellOverride ? { shellOverride } : {}), ...(projectRuntime ? { projectRuntime } : {}), diff --git a/src/renderer/src/i18n/locales/de.json b/src/renderer/src/i18n/locales/de.json index c1ff4d55106..89f3b21e9dc 100644 --- a/src/renderer/src/i18n/locales/de.json +++ b/src/renderer/src/i18n/locales/de.json @@ -11307,7 +11307,7 @@ }, "agent-workspace-trust-copy": { "title": "Dem Ordner vertrauen, wenn Orca einen Agenten startet", - "description": "Von Orca gestartete Agenten können die Abfrage „Do you trust this folder?“ überspringen, sodass die Agenten-Hooks und Einstellungen des Projekts sofort ausgeführt werden. Orca fragt einmal pro Repository beim ersten Agentenstart dort nach; die Antwort lässt sich in den Einstellungen dieses Repositorys ändern. Ausgeschaltet fragen Agenten immer selbst nach. Ein Agent, der unbeaufsichtigt läuft und nachfragt, etwa ein Orchestrierungs-Worker, eine Automatisierung oder ein vom Smartphone gestarteter Agent, wartet, bis die Frage in seinem Terminal beantwortet wird. Bereits als vertrauenswürdig eingestufte Ordner bleiben vertrauenswürdig." + "description": "Von Orca gestartete Agenten können die Abfrage „Do you trust this folder?“ überspringen, sodass die Agenten-Hooks und Einstellungen des Projekts sofort ausgeführt werden. In den Einstellungen eines Repositorys lässt sich das für dieses Repository ausschalten. Ausgeschaltet fragen Agenten immer selbst nach. Ein Agent, der unbeaufsichtigt läuft und nachfragt, etwa ein Orchestrierungs-Worker, eine Automatisierung oder ein vom Smartphone gestarteter Agent, wartet, bis die Frage in seinem Terminal beantwortet wird. Bereits als vertrauenswürdig eingestufte Ordner bleiben vertrauenswürdig." }, "agent-generated-tab-title-copy": { "19ad21615a": "Tab-Titel automatisch generieren", @@ -18987,14 +18987,10 @@ } }, "agent-workspace-trust-decision": { - "prompt-title": "Dem Repository {{repo}} für Agenten vertrauen?", - "prompt-description": "Agenten wie Claude Code und Codex fragen „Do you trust this folder?“, bevor sie die eigenen Agenten-Einstellungen und Hooks des Projekts laden, die Befehle ausführen können. Orca kann diese Frage in jedem Worktree dieses Repositorys automatisch mit Ja beantworten.", - "prompt-remembered": "Orca merkt sich die Antwort für dieses Repository. Sie lässt sich in den Einstellungen des Repositorys ändern.", "trust-repository": "Dem Repository vertrauen", "let-agents-ask": "Agenten fragen lassen", "setting-title": "Ordnervertrauen für Agenten", "setting-description": "Legt fest, ob von Orca gestartete Agenten in den Worktrees dieses Repositorys die Abfrage „Do you trust this folder?“ überspringen und die Agenten-Einstellungen und Hooks des Projekts sofort laden.", - "setting-undecided": "Noch nicht entschieden. Orca fragt beim nächsten Agentenstart hier nach.", "setting-global-off": "Das Ordnervertrauen ist in den Agenten-Einstellungen ausgeschaltet, daher fragen Agenten immer nach.", "keyword-trust": "vertrauen", "keyword-hooks": "Hooks" diff --git a/src/renderer/src/i18n/locales/en.json b/src/renderer/src/i18n/locales/en.json index 6b5b44dcb92..c224064e058 100644 --- a/src/renderer/src/i18n/locales/en.json +++ b/src/renderer/src/i18n/locales/en.json @@ -11768,7 +11768,7 @@ }, "agent-workspace-trust-copy": { "title": "Trust the folder when Orca starts an agent", - "description": "Agents Orca starts can skip their \"Do you trust this folder?\" prompt, so the project's agent hooks and settings run right away. Orca asks once per repository the first time you start an agent there, and you can change the answer in that repository's settings. Turn this off to let agents always ask. An agent that asks while running without you watching, such as an orchestration worker, an automation or an agent started from your phone, waits until you answer in its terminal. Folders already trusted stay trusted." + "description": "Agents Orca starts can skip their \"Do you trust this folder?\" prompt, so the project's agent hooks and settings run right away. You can turn this off for one repository in that repository's settings. Turn this off to let agents always ask. An agent that asks while running without you watching, such as an orchestration worker, an automation or an agent started from your phone, waits until you answer in its terminal. Folders already trusted stay trusted." }, "agent-generated-tab-title-copy": { "19ad21615a": "Auto-generate tab titles", @@ -19607,14 +19607,10 @@ } }, "agent-workspace-trust-decision": { - "prompt-title": "Trust {{repo}} for agents?", - "prompt-description": "Agents like Claude Code and Codex ask \"Do you trust this folder?\" before they load the project's own agent settings and hooks, which can run commands. Orca can answer yes for you in every worktree of this repository.", - "prompt-remembered": "Orca remembers your answer for this repository. You can change it in the repository's settings.", "trust-repository": "Trust repository", "let-agents-ask": "Let agents ask", "setting-title": "Agent Folder Trust", "setting-description": "Whether agents Orca starts in this repository's worktrees skip their \"Do you trust this folder?\" prompt and load the project's agent settings and hooks right away.", - "setting-undecided": "Not decided yet. Orca will ask the next time you start an agent here.", "setting-global-off": "Folder trust is turned off in Agents settings, so agents always ask.", "keyword-trust": "trust", "keyword-hooks": "hooks" diff --git a/src/renderer/src/i18n/locales/es.json b/src/renderer/src/i18n/locales/es.json index ccf9c77f2ad..1ad7f1e33dc 100644 --- a/src/renderer/src/i18n/locales/es.json +++ b/src/renderer/src/i18n/locales/es.json @@ -10905,7 +10905,7 @@ }, "agent-workspace-trust-copy": { "title": "Confiar en la carpeta cuando Orca inicia un agente", - "description": "Los agentes que inicia Orca pueden omitir su pregunta \"Do you trust this folder?\", así que los hooks y ajustes de agente del proyecto se ejecutan de inmediato. Orca pregunta una vez por repositorio la primera vez que inicias un agente allí, y puedes cambiar la respuesta en los ajustes de ese repositorio. Desactívalo para que los agentes pregunten siempre. Un agente que pregunta mientras se ejecuta sin que estés mirando, como un worker de orquestación, una automatización o un agente iniciado desde tu teléfono, espera hasta que respondas en su terminal. Las carpetas que ya son de confianza lo siguen siendo." + "description": "Los agentes que inicia Orca pueden omitir su pregunta \"Do you trust this folder?\", así que los hooks y ajustes de agente del proyecto se ejecutan de inmediato. Puedes desactivarlo para un repositorio en los ajustes de ese repositorio. Desactívalo para que los agentes pregunten siempre. Un agente que pregunta mientras se ejecuta sin que estés mirando, como un worker de orquestación, una automatización o un agente iniciado desde tu teléfono, espera hasta que respondas en su terminal. Las carpetas que ya son de confianza lo siguen siendo." }, "agent-generated-tab-title-copy": { "19ad21615a": "Generar títulos de pestañas automáticamente", @@ -18781,14 +18781,10 @@ } }, "agent-workspace-trust-decision": { - "prompt-title": "¿Confiar en {{repo}} para los agentes?", - "prompt-description": "Los agentes como Claude Code y Codex preguntan \"Do you trust this folder?\" antes de cargar los ajustes y hooks de agente propios del proyecto, que pueden ejecutar comandos. Orca puede responder que sí por ti en todos los worktrees de este repositorio.", - "prompt-remembered": "Orca recuerda tu respuesta para este repositorio. Puedes cambiarla en los ajustes del repositorio.", "trust-repository": "Confiar en el repositorio", "let-agents-ask": "Dejar que los agentes pregunten", "setting-title": "Confianza de carpetas para agentes", "setting-description": "Indica si los agentes que Orca inicia en los worktrees de este repositorio omiten su pregunta \"Do you trust this folder?\" y cargan de inmediato los ajustes y hooks de agente del proyecto.", - "setting-undecided": "Aún no se ha decidido. Orca preguntará la próxima vez que inicies un agente aquí.", "setting-global-off": "La confianza de carpetas está desactivada en los ajustes de Agentes, así que los agentes siempre preguntan.", "keyword-trust": "confianza", "keyword-hooks": "hooks" diff --git a/src/renderer/src/i18n/locales/fr.json b/src/renderer/src/i18n/locales/fr.json index 11bcd049e30..c88d0253779 100644 --- a/src/renderer/src/i18n/locales/fr.json +++ b/src/renderer/src/i18n/locales/fr.json @@ -11175,7 +11175,7 @@ }, "agent-workspace-trust-copy": { "title": "Faire confiance au dossier quand Orca lance un agent", - "description": "Les agents lancés par Orca peuvent ignorer leur question « Do you trust this folder? » : les hooks et paramètres d'agent du projet s'exécutent donc tout de suite. Orca pose la question une fois par dépôt, la première fois que vous y lancez un agent, et vous pouvez modifier la réponse dans les paramètres de ce dépôt. Désactivez cette option pour que les agents posent toujours la question. Un agent qui pose la question alors qu'il tourne sans que vous le surveilliez, comme un worker d'orchestration, une automatisation ou un agent lancé depuis votre téléphone, attend que vous répondiez dans son terminal. Les dossiers déjà approuvés le restent." + "description": "Les agents lancés par Orca peuvent ignorer leur question « Do you trust this folder? » : les hooks et paramètres d'agent du projet s'exécutent donc tout de suite. Vous pouvez désactiver cette option pour un dépôt dans les paramètres de ce dépôt. Désactivez cette option pour que les agents posent toujours la question. Un agent qui pose la question alors qu'il tourne sans que vous le surveilliez, comme un worker d'orchestration, une automatisation ou un agent lancé depuis votre téléphone, attend que vous répondiez dans son terminal. Les dossiers déjà approuvés le restent." }, "agent-generated-tab-title-copy": { "19ad21615a": "Générer automatiquement les titres des onglets", @@ -19020,14 +19020,10 @@ } }, "agent-workspace-trust-decision": { - "prompt-title": "Faire confiance à {{repo}} pour les agents ?", - "prompt-description": "Les agents comme Claude Code et Codex demandent « Do you trust this folder? » avant de charger les paramètres et hooks d'agent propres au projet, qui peuvent exécuter des commandes. Orca peut répondre oui à votre place dans chaque worktree de ce dépôt.", - "prompt-remembered": "Orca mémorise votre réponse pour ce dépôt. Vous pouvez la modifier dans les paramètres du dépôt.", "trust-repository": "Faire confiance au dépôt", "let-agents-ask": "Laisser les agents demander", "setting-title": "Confiance des dossiers pour les agents", "setting-description": "Indique si les agents lancés par Orca dans les worktrees de ce dépôt ignorent leur question « Do you trust this folder? » et chargent tout de suite les paramètres et hooks d'agent du projet.", - "setting-undecided": "Pas encore décidé. Orca posera la question la prochaine fois que vous lancerez un agent ici.", "setting-global-off": "La confiance des dossiers est désactivée dans les paramètres Agents : les agents posent donc toujours la question.", "keyword-trust": "confiance", "keyword-hooks": "hooks" diff --git a/src/renderer/src/i18n/locales/ja.json b/src/renderer/src/i18n/locales/ja.json index 6d0e225625c..8dd84882a81 100644 --- a/src/renderer/src/i18n/locales/ja.json +++ b/src/renderer/src/i18n/locales/ja.json @@ -10664,7 +10664,7 @@ }, "agent-workspace-trust-copy": { "title": "Orca が Agent を起動するときにフォルダーを信頼する", - "description": "Orca が起動した Agent は「Do you trust this folder?」の確認を省略できるため、プロジェクトの Agent のフックと設定がすぐに実行されます。Orca はリポジトリごとに、初めて Agent を起動したときに一度だけ確認します。回答はそのリポジトリの設定で変更できます。オフにすると Agent は常に確認します。オーケストレーションのワーカー、オートメーション、スマートフォンから起動した Agent など、あなたが見ていない間に動く Agent が確認を求めた場合は、そのターミナルで応答するまで待機します。すでに信頼済みのフォルダーは信頼されたままです。" + "description": "Orca が起動した Agent は「Do you trust this folder?」の確認を省略できるため、プロジェクトの Agent のフックと設定がすぐに実行されます。リポジトリごとに、そのリポジトリの設定でオフにできます。オフにすると Agent は常に確認します。オーケストレーションのワーカー、オートメーション、スマートフォンから起動した Agent など、あなたが見ていない間に動く Agent が確認を求めた場合は、そのターミナルで応答するまで待機します。すでに信頼済みのフォルダーは信頼されたままです。" }, "agent-generated-tab-title-copy": { "19ad21615a": "タブタイトルを自動生成", @@ -18308,14 +18308,10 @@ } }, "agent-workspace-trust-decision": { - "prompt-title": "{{repo}} を Agent 用に信頼しますか?", - "prompt-description": "Claude Code や Codex などの Agent は、プロジェクト独自の Agent 設定とフック(コマンドを実行できます)を読み込む前に「Do you trust this folder?」と確認します。Orca がこのリポジトリのすべてのワークツリーで代わりに「はい」と答えることができます。", - "prompt-remembered": "Orca はこのリポジトリへの回答を記憶します。リポジトリの設定で変更できます。", "trust-repository": "リポジトリを信頼", "let-agents-ask": "Agent に確認させる", "setting-title": "Agent のフォルダー信頼", "setting-description": "このリポジトリのワークツリーで Orca が起動した Agent が「Do you trust this folder?」の確認を省略し、プロジェクトの Agent 設定とフックをすぐに読み込むかどうかを指定します。", - "setting-undecided": "まだ決めていません。次にここで Agent を起動したときに Orca が確認します。", "setting-global-off": "Agent 設定でフォルダーの信頼がオフになっているため、Agent は常に確認します。", "keyword-trust": "信頼", "keyword-hooks": "フック" diff --git a/src/renderer/src/i18n/locales/ko.json b/src/renderer/src/i18n/locales/ko.json index 1bc9f6dea18..15bab956dc6 100644 --- a/src/renderer/src/i18n/locales/ko.json +++ b/src/renderer/src/i18n/locales/ko.json @@ -10627,7 +10627,7 @@ }, "agent-workspace-trust-copy": { "title": "Orca가 에이전트를 시작할 때 폴더 신뢰", - "description": "Orca가 시작한 에이전트는 \"Do you trust this folder?\" 확인을 건너뛸 수 있으므로, 프로젝트의 에이전트 훅과 설정이 바로 실행됩니다. Orca는 저장소마다 처음 에이전트를 시작할 때 한 번 묻고, 답변은 해당 저장소 설정에서 변경할 수 있습니다. 이 설정을 끄면 에이전트가 항상 확인을 요청합니다. 오케스트레이션 워커, 자동화, 휴대폰에서 시작한 에이전트처럼 사용자가 지켜보지 않는 동안 실행되는 에이전트가 확인을 요청하면 해당 터미널에서 응답할 때까지 기다립니다. 이미 신뢰한 폴더는 계속 신뢰됩니다." + "description": "Orca가 시작한 에이전트는 \"Do you trust this folder?\" 확인을 건너뛸 수 있으므로, 프로젝트의 에이전트 훅과 설정이 바로 실행됩니다. 저장소별로 해당 저장소 설정에서 끌 수 있습니다. 이 설정을 끄면 에이전트가 항상 확인을 요청합니다. 오케스트레이션 워커, 자동화, 휴대폰에서 시작한 에이전트처럼 사용자가 지켜보지 않는 동안 실행되는 에이전트가 확인을 요청하면 해당 터미널에서 응답할 때까지 기다립니다. 이미 신뢰한 폴더는 계속 신뢰됩니다." }, "agent-generated-tab-title-copy": { "19ad21615a": "탭 제목 자동 생성", @@ -18308,14 +18308,10 @@ } }, "agent-workspace-trust-decision": { - "prompt-title": "{{repo}}을(를) 에이전트용으로 신뢰할까요?", - "prompt-description": "Claude Code, Codex 같은 에이전트는 명령을 실행할 수 있는 프로젝트 자체의 에이전트 설정과 훅을 불러오기 전에 \"Do you trust this folder?\"라고 묻습니다. Orca가 이 저장소의 모든 작업 트리에서 대신 \"예\"라고 답할 수 있습니다.", - "prompt-remembered": "Orca는 이 저장소에 대한 답변을 기억합니다. 저장소 설정에서 변경할 수 있습니다.", "trust-repository": "저장소 신뢰", "let-agents-ask": "에이전트가 묻도록 하기", "setting-title": "에이전트 폴더 신뢰", "setting-description": "이 저장소의 작업 트리에서 Orca가 시작한 에이전트가 \"Do you trust this folder?\" 확인을 건너뛰고 프로젝트의 에이전트 설정과 훅을 바로 불러올지 정합니다.", - "setting-undecided": "아직 정하지 않았습니다. 다음에 여기서 에이전트를 시작하면 Orca가 묻습니다.", "setting-global-off": "에이전트 설정에서 폴더 신뢰가 꺼져 있으므로 에이전트가 항상 확인을 요청합니다.", "keyword-trust": "신뢰", "keyword-hooks": "훅" diff --git a/src/renderer/src/i18n/locales/pt.json b/src/renderer/src/i18n/locales/pt.json index d6a9f26f7dc..8200201faf8 100644 --- a/src/renderer/src/i18n/locales/pt.json +++ b/src/renderer/src/i18n/locales/pt.json @@ -8716,7 +8716,7 @@ }, "agent-workspace-trust-copy": { "title": "Confiar na pasta quando o Orca iniciar um agente", - "description": "Os agentes iniciados pelo Orca podem pular a pergunta “Você confia nesta pasta?”, para que os hooks e configurações de agentes do projeto sejam executados imediatamente. O Orca pergunta uma vez por repositório, na primeira vez que você inicia um agente nele, e você pode alterar a resposta nas configurações desse repositório. Desative para que os agentes sempre perguntem. Um agente que pergunta enquanto é executado sem sua supervisão, como um worker de orquestração, uma automação ou um agente iniciado pelo celular, espera até você responder no terminal dele. As pastas já confiáveis continuam confiáveis." + "description": "Os agentes iniciados pelo Orca podem pular a pergunta “Você confia nesta pasta?”, para que os hooks e configurações de agentes do projeto sejam executados imediatamente. Você pode desativar isso para um repositório nas configurações desse repositório. Desative para que os agentes sempre perguntem. Um agente que pergunta enquanto é executado sem sua supervisão, como um worker de orquestração, uma automação ou um agente iniciado pelo celular, espera até você responder no terminal dele. As pastas já confiáveis continuam confiáveis." }, "MachineNameField": { "label": "Nome da máquina", @@ -20016,14 +20016,10 @@ } }, "agent-workspace-trust-decision": { - "prompt-title": "Confiar em {{repo}} para agentes?", - "prompt-description": "Agentes como Claude Code e Codex perguntam “Você confia nesta pasta?” antes de carregar as configurações e hooks de agentes do próprio projeto, que podem executar comandos. O Orca pode responder sim por você em todas as worktrees deste repositório.", - "prompt-remembered": "O Orca lembra sua resposta para este repositório. Você pode alterá-la nas configurações do repositório.", "trust-repository": "Confiar no repositório", "let-agents-ask": "Deixar os agentes perguntarem", "setting-title": "Confiança de Pasta para Agentes", "setting-description": "Define se os agentes que o Orca inicia nas worktrees deste repositório pulam a pergunta “Você confia nesta pasta?” e carregam imediatamente as configurações e hooks de agentes do projeto.", - "setting-undecided": "Ainda não decidido. O Orca vai perguntar na próxima vez que você iniciar um agente aqui.", "setting-global-off": "A confiança de pasta está desativada nas configurações de Agentes, então os agentes sempre perguntam.", "keyword-trust": "confiança", "keyword-hooks": "hooks" diff --git a/src/renderer/src/i18n/locales/ru.json b/src/renderer/src/i18n/locales/ru.json index 263afc7c280..636209ed965 100644 --- a/src/renderer/src/i18n/locales/ru.json +++ b/src/renderer/src/i18n/locales/ru.json @@ -11659,7 +11659,7 @@ }, "agent-workspace-trust-copy": { "title": "Доверять папке при запуске агента из Orca", - "description": "Агенты, запущенные Orca, могут пропускать вопрос «Доверяете ли вы этой папке?», поэтому хуки и настройки агентов проекта выполняются сразу. Orca спрашивает один раз для каждого репозитория при первом запуске агента в нём; ответ можно изменить в настройках этого репозитория. Отключите, чтобы агенты всегда спрашивали сами. Агент, который задаёт этот вопрос, работая без вашего наблюдения, например исполнитель оркестрации, автоматизация или агент, запущенный с телефона, ждёт ответа в своём терминале. Уже доверенные папки сохраняют доверие." + "description": "Агенты, запущенные Orca, могут пропускать вопрос «Доверяете ли вы этой папке?», поэтому хуки и настройки агентов проекта выполняются сразу. Для отдельного репозитория это можно отключить в его настройках. Отключите, чтобы агенты всегда спрашивали сами. Агент, который задаёт этот вопрос, работая без вашего наблюдения, например исполнитель оркестрации, автоматизация или агент, запущенный с телефона, ждёт ответа в своём терминале. Уже доверенные папки сохраняют доверие." }, "agent-generated-tab-title-copy": { "19ad21615a": "Автоматически создавать заголовки вкладок", @@ -19557,14 +19557,10 @@ } }, "agent-workspace-trust-decision": { - "prompt-title": "Доверять {{repo}} для агентов?", - "prompt-description": "Агенты, такие как Claude Code и Codex, спрашивают «Доверяете ли вы этой папке?», прежде чем загрузить собственные настройки и хуки агентов проекта, которые могут выполнять команды. Orca может отвечать «да» за вас во всех рабочих деревьях этого репозитория.", - "prompt-remembered": "Orca запоминает ваш ответ для этого репозитория. Его можно изменить в настройках репозитория.", "trust-repository": "Доверять репозиторию", "let-agents-ask": "Пусть агенты спрашивают", "setting-title": "Доверие к папкам для агентов", "setting-description": "Пропускают ли агенты, запущенные Orca в рабочих деревьях этого репозитория, вопрос «Доверяете ли вы этой папке?» и сразу загружают настройки и хуки агентов проекта.", - "setting-undecided": "Решение ещё не принято. Orca спросит, когда вы в следующий раз запустите здесь агента.", "setting-global-off": "Доверие к папкам отключено в настройках агентов, поэтому агенты всегда спрашивают.", "keyword-trust": "доверие", "keyword-hooks": "хуки" diff --git a/src/renderer/src/i18n/locales/tr.json b/src/renderer/src/i18n/locales/tr.json index 71c4f37cc42..61b341f799d 100644 --- a/src/renderer/src/i18n/locales/tr.json +++ b/src/renderer/src/i18n/locales/tr.json @@ -11291,7 +11291,7 @@ }, "agent-workspace-trust-copy": { "title": "Orca bir ajan başlattığında klasöre güven", - "description": "Orca'nın başlattığı ajanlar \"Bu klasöre güveniyor musunuz?\" sorusunu atlayabilir; böylece projenin ajan kancaları ve ayarları hemen çalışır. Orca her repository için orada ilk kez bir ajan başlattığınızda bir kez sorar; yanıtı o repository'nin ayarlarından değiştirebilirsiniz. Ajanların her zaman sorması için bunu kapatın. Orkestrasyon çalışanı, otomasyon veya telefonunuzdan başlatılan bir ajan gibi gözetiminiz dışında çalışırken soru soran bir ajan, terminalinde yanıt verene kadar bekler. Zaten güvenilen klasörlere güvenilmeye devam edilir." + "description": "Orca'nın başlattığı ajanlar \"Bu klasöre güveniyor musunuz?\" sorusunu atlayabilir; böylece projenin ajan kancaları ve ayarları hemen çalışır. Bunu bir repository için o repository'nin ayarlarından kapatabilirsiniz. Ajanların her zaman sorması için bunu kapatın. Orkestrasyon çalışanı, otomasyon veya telefonunuzdan başlatılan bir ajan gibi gözetiminiz dışında çalışırken soru soran bir ajan, terminalinde yanıt verene kadar bekler. Zaten güvenilen klasörlere güvenilmeye devam edilir." }, "agent-generated-tab-title-copy": { "19ad21615a": "Sekme başlıklarını otomatik oluştur", @@ -18971,14 +18971,10 @@ } }, "agent-workspace-trust-decision": { - "prompt-title": "{{repo}} ajanlar için güvenilir olsun mu?", - "prompt-description": "Claude Code ve Codex gibi ajanlar, projenin komut çalıştırabilen kendi ajan ayarlarını ve kancalarını yüklemeden önce \"Bu klasöre güveniyor musunuz?\" diye sorar. Orca, bu repository'nin tüm worktree'lerinde sizin yerinize evet diyebilir.", - "prompt-remembered": "Orca bu repository için verdiğiniz yanıtı hatırlar. Yanıtı repository ayarlarından değiştirebilirsiniz.", "trust-repository": "Repository'ye güven", "let-agents-ask": "Ajanlar sorsun", "setting-title": "Ajan Klasör Güveni", "setting-description": "Orca'nın bu repository'nin worktree'lerinde başlattığı ajanların \"Bu klasöre güveniyor musunuz?\" sorusunu atlayıp projenin ajan ayarlarını ve kancalarını hemen yükleyip yüklemeyeceği.", - "setting-undecided": "Henüz karar verilmedi. Burada bir sonraki ajan başlatışınızda Orca soracak.", "setting-global-off": "Klasör güveni Ajanlar ayarlarında kapalı olduğundan ajanlar her zaman sorar.", "keyword-trust": "güven", "keyword-hooks": "kancalar" diff --git a/src/renderer/src/i18n/locales/zh-Hant.json b/src/renderer/src/i18n/locales/zh-Hant.json index 68e7f0ba2ae..477b8c9fbf0 100644 --- a/src/renderer/src/i18n/locales/zh-Hant.json +++ b/src/renderer/src/i18n/locales/zh-Hant.json @@ -11128,7 +11128,7 @@ }, "agent-workspace-trust-copy": { "title": "Orca 啟動代理程式時信任該資料夾", - "description": "Orca 啟動的代理程式可以略過「是否信任此資料夾?」的詢問,因此專案的代理程式鉤子和設定會立即執行。Orca 會在每個存放庫中首次啟動代理程式時詢問一次,你可以在該存放庫的設定中變更回答。關閉此項可讓代理程式一律自行詢問。在你沒有看著時執行的代理程式(例如編排工作行程、自動化以及從手機啟動的代理程式)若發出詢問,會一直等待,直到你在它的終端機中作答。已信任的資料夾仍保持信任。" + "description": "Orca 啟動的代理程式可以略過「是否信任此資料夾?」的詢問,因此專案的代理程式鉤子和設定會立即執行。你可以在某個存放庫的設定中為該存放庫關閉此項。關閉此項可讓代理程式一律自行詢問。在你沒有看著時執行的代理程式(例如編排工作行程、自動化以及從手機啟動的代理程式)若發出詢問,會一直等待,直到你在它的終端機中作答。已信任的資料夾仍保持信任。" }, "agent-generated-tab-title-copy": { "19ad21615a": "自動產生分頁標題", @@ -18726,14 +18726,10 @@ } }, "agent-workspace-trust-decision": { - "prompt-title": "要為代理程式信任 {{repo}} 嗎?", - "prompt-description": "Claude Code 和 Codex 等代理程式在載入專案自帶的代理程式設定和鉤子(可以執行命令)之前,會詢問「是否信任此資料夾?」。Orca 可以在此存放庫的每個工作樹中替你回答「是」。", - "prompt-remembered": "Orca 會記住你對此存放庫的回答。你可以在存放庫設定中變更。", "trust-repository": "信任存放庫", "let-agents-ask": "讓代理程式詢問", "setting-title": "代理程式資料夾信任", "setting-description": "Orca 在此存放庫的工作樹中啟動的代理程式是否略過「是否信任此資料夾?」的詢問,並立即載入專案的代理程式設定和鉤子。", - "setting-undecided": "尚未決定。下次你在這裡啟動代理程式時,Orca 會詢問。", "setting-global-off": "已在代理程式設定中關閉資料夾信任,因此代理程式一律會詢問。", "keyword-trust": "信任", "keyword-hooks": "鉤子" diff --git a/src/renderer/src/i18n/locales/zh.json b/src/renderer/src/i18n/locales/zh.json index a2a9b61bbe6..f65e600cc9d 100644 --- a/src/renderer/src/i18n/locales/zh.json +++ b/src/renderer/src/i18n/locales/zh.json @@ -10627,7 +10627,7 @@ }, "agent-workspace-trust-copy": { "title": "Orca 启动智能体时信任该文件夹", - "description": "Orca 启动的智能体可以跳过“Do you trust this folder?”提示,因此项目的智能体钩子和设置会立即运行。Orca 会在每个仓库中首次启动智能体时询问一次,你可以在该仓库的设置中更改回答。关闭此项可让智能体始终自行询问。在你没有看着时运行的智能体(例如编排工作进程、自动化以及从手机启动的智能体)如果发出询问,会一直等待,直到你在它的终端中作答。已信任的文件夹仍保持信任。" + "description": "Orca 启动的智能体可以跳过“Do you trust this folder?”提示,因此项目的智能体钩子和设置会立即运行。你可以在某个仓库的设置中为该仓库关闭此项。关闭此项可让智能体始终自行询问。在你没有看着时运行的智能体(例如编排工作进程、自动化以及从手机启动的智能体)如果发出询问,会一直等待,直到你在它的终端中作答。已信任的文件夹仍保持信任。" }, "agent-generated-tab-title-copy": { "19ad21615a": "自动生成标签页标题", @@ -18308,14 +18308,10 @@ } }, "agent-workspace-trust-decision": { - "prompt-title": "为智能体信任 {{repo}}?", - "prompt-description": "Claude Code 和 Codex 等智能体在加载项目自带的智能体设置和钩子(可以运行命令)之前,会询问“Do you trust this folder?”。Orca 可以在此仓库的每个工作树中替你回答“是”。", - "prompt-remembered": "Orca 会记住你对此仓库的回答。你可以在仓库设置中更改。", "trust-repository": "信任仓库", "let-agents-ask": "让智能体询问", "setting-title": "智能体文件夹信任", "setting-description": "Orca 在此仓库的工作树中启动的智能体是否跳过“Do you trust this folder?”提示,并立即加载项目的智能体设置和钩子。", - "setting-undecided": "尚未决定。下次你在这里启动智能体时,Orca 会询问。", "setting-global-off": "已在智能体设置中关闭文件夹信任,因此智能体总会询问。", "keyword-trust": "信任", "keyword-hooks": "钩子" diff --git a/src/renderer/src/lib/agent-session/launch/parity/launch-parity-renderer.test-fixture.ts b/src/renderer/src/lib/agent-session/launch/parity/launch-parity-renderer.test-fixture.ts index fb9a81e77d4..8c3cb21df18 100644 --- a/src/renderer/src/lib/agent-session/launch/parity/launch-parity-renderer.test-fixture.ts +++ b/src/renderer/src/lib/agent-session/launch/parity/launch-parity-renderer.test-fixture.ts @@ -81,9 +81,7 @@ export function launchWorkspaceState( addedAt: 0, kind: 'git', connectionId: null, - executionHostId: 'local', - // Parity rows model a repo whose first-launch trust answer is already stored. - agentWorkspaceTrust: 'pre-trust' + executionHostId: 'local' } : null const worktree = { diff --git a/src/shared/agent-launch/permissions/agent-workspace-trust-decision.ts b/src/shared/agent-launch/permissions/agent-workspace-trust-decision.ts index fc7aefaf38e..31819184293 100644 --- a/src/shared/agent-launch/permissions/agent-workspace-trust-decision.ts +++ b/src/shared/agent-launch/permissions/agent-workspace-trust-decision.ts @@ -1,5 +1,5 @@ export type AgentWorkspaceTrustSettings = { - /** Off: never pre-trust. On: pre-trust per each repo's decision (asked on first desktop launch). Defaults on. */ + /** Off: never pre-trust. On: pre-trust unless the repo opted out. Defaults on. */ agentWorkspaceTrustEnabled: boolean /** One-shot guard: repos saved before per-repo trust decisions were pinned to pre-trust. */ agentWorkspaceTrustExistingReposMigrated?: boolean @@ -14,7 +14,7 @@ export function sanitizeAgentWorkspaceTrustDecision( return value === 'pre-trust' || value === 'agent-asks' ? value : undefined } -/** Global off always wins; an undecided repo keeps the pre-existing pre-trust default so unattended launches never stall. */ +/** Global off always wins; a repo is pre-trusted unless it opted out. */ export function shouldPreTrustAgentWorkspace( globalEnabled: boolean | undefined, repoDecision: AgentWorkspaceTrustDecision | undefined @@ -22,14 +22,6 @@ export function shouldPreTrustAgentWorkspace( return globalEnabled !== false && repoDecision !== 'agent-asks' } -/** Only a launch the user is watching asks; it must settle before the spawn reads the answer. */ -export function needsAgentWorkspaceTrustDecision( - globalEnabled: boolean | undefined, - repoDecision: AgentWorkspaceTrustDecision | undefined -): boolean { - return globalEnabled !== false && repoDecision === undefined -} - export function resolveAgentWorkspaceTrustOwner(args: { worktreeId?: string trustWorktreeId?: string diff --git a/src/shared/plugins/languages/protected-copy-keys.json b/src/shared/plugins/languages/protected-copy-keys.json index df88ae07257..12adc04a91b 100644 --- a/src/shared/plugins/languages/protected-copy-keys.json +++ b/src/shared/plugins/languages/protected-copy-keys.json @@ -70,13 +70,9 @@ "auto.components.agent-workspace-trust-decision.keyword-hooks", "auto.components.agent-workspace-trust-decision.keyword-trust", "auto.components.agent-workspace-trust-decision.let-agents-ask", - "auto.components.agent-workspace-trust-decision.prompt-description", - "auto.components.agent-workspace-trust-decision.prompt-remembered", - "auto.components.agent-workspace-trust-decision.prompt-title", "auto.components.agent-workspace-trust-decision.setting-description", "auto.components.agent-workspace-trust-decision.setting-global-off", "auto.components.agent-workspace-trust-decision.setting-title", - "auto.components.agent-workspace-trust-decision.setting-undecided", "auto.components.agent-workspace-trust-decision.trust-repository", "auto.components.automations.AutomationsPage.080dcb5fbb", "auto.components.automations.AutomationsPage.15e0bfb13b", diff --git a/tests/e2e/agent-workspace-trust-first-launch.spec.ts b/tests/e2e/agent-workspace-trust-first-launch.spec.ts deleted file mode 100644 index da4d356cd12..00000000000 --- a/tests/e2e/agent-workspace-trust-first-launch.spec.ts +++ /dev/null @@ -1,84 +0,0 @@ -import type { Page } from '@stablyai/playwright-test' -import { expect, test } from './helpers/orca-app' -import { - configureGoldenStubAgent, - getGoldenStubAgentLaunchEnv, - GOLDEN_STUB_READY_MARKER -} from './helpers/golden-stub-agent' -import { addRepoWithWorkspaces } from './helpers/repo-with-workspaces' -import { ensureTerminalVisible, switchToWorktree, waitForSessionReady } from './helpers/store' -import { focusActiveTerminalInput, waitForTerminalOutput } from './helpers/terminal' - -test.use({ launchEnv: getGoldenStubAgentLaunchEnv() }) - -const CLAUDE_MENU_ITEM = /^Claude(?:\s|$)/i - -async function startClaudeFromNewTab(page: Page): Promise { - await page.getByRole('button', { name: 'New tab' }).click({ force: true }) - const launchOption = page.getByRole('menuitem', { name: CLAUDE_MENU_ITEM }).first() - await expect(launchOption).toBeVisible({ timeout: 15_000 }) - await launchOption.click({ force: true }) -} - -async function waitForAgentReady(page: Page): Promise { - await focusActiveTerminalInput(page) - await waitForTerminalOutput(page, GOLDEN_STUB_READY_MARKER, 20_000) -} - -async function readRepoDecision(page: Page, repoId: string): Promise { - return page.evaluate( - (id) => - window.__store?.getState().repos.find((repo) => repo.id === id)?.agentWorkspaceTrust ?? null, - repoId - ) -} - -test('the first agent launch in a new repo asks once, and its worktrees share the answer', async ({ - orcaPage -}, testInfo) => { - await waitForSessionReady(orcaPage) - await configureGoldenStubAgent(orcaPage, { agent: 'claude' }) - // A repo added now has no stored answer, like a fresh clone after the upgrade. - const { repoId, mainId, ids } = await addRepoWithWorkspaces(orcaPage, testInfo, 1) - await switchToWorktree(orcaPage, mainId) - await ensureTerminalVisible(orcaPage) - - await startClaudeFromNewTab(orcaPage) - const prompt = orcaPage.getByRole('dialog', { name: /for agents\?$/ }) - await expect(prompt).toBeVisible({ timeout: 15_000 }) - await orcaPage.screenshot({ path: testInfo.outputPath('first-launch-prompt.png') }) - expect(await readRepoDecision(orcaPage, repoId)).toBeNull() - - await prompt.getByRole('button', { name: 'Let agents ask', exact: true }).click() - await expect(prompt).toBeHidden() - await waitForAgentReady(orcaPage) - expect(await readRepoDecision(orcaPage, repoId)).toBe('agent-asks') - - // Another worktree of the same repo reuses the stored answer. - const [sibling] = ids - if (!sibling) { - throw new Error('the repo needs a second worktree') - } - await switchToWorktree(orcaPage, sibling) - await ensureTerminalVisible(orcaPage) - await startClaudeFromNewTab(orcaPage) - await waitForAgentReady(orcaPage) - await expect(prompt).toHaveCount(0) - - await orcaPage.evaluate((id) => { - const state = window.__store?.getState() - state?.setSettingsSearchQuery('') - state?.openSettingsTarget({ pane: 'repo', repoId: id }) - state?.openSettingsPage() - }, repoId) - const trustControl = orcaPage.getByRole('radiogroup', { name: 'Agent Folder Trust' }) - await trustControl.scrollIntoViewIfNeeded() - await expect(trustControl.getByRole('radio', { name: 'Let agents ask' })).toHaveAttribute( - 'aria-checked', - 'true' - ) - await orcaPage.screenshot({ path: testInfo.outputPath('repo-setting.png') }) - - await trustControl.getByRole('radio', { name: 'Trust repository' }).click() - await expect.poll(() => readRepoDecision(orcaPage, repoId)).toBe('pre-trust') -}) diff --git a/tests/e2e/helpers/orca-app.ts b/tests/e2e/helpers/orca-app.ts index 32b8e2b8ab5..81d31f1959d 100644 --- a/tests/e2e/helpers/orca-app.ts +++ b/tests/e2e/helpers/orca-app.ts @@ -259,12 +259,8 @@ export const test = base.extend({ if (!repo) { return false } - // Show external worktrees (hidden by default); pre-answer the first-launch trust prompt. - const show = { - externalWorktreeVisibility: 'show', - agentWorkspaceTrust: 'pre-trust' - } as const - await store.getState().updateRepo(repo.id, show) + // Show the fixture’s external Git worktrees, which newly added repos hide by default. + await store.getState().updateRepo(repo.id, { externalWorktreeVisibility: 'show' }) return true }, seededRepoId), { diff --git a/tests/e2e/helpers/orca-restart.ts b/tests/e2e/helpers/orca-restart.ts index 6f1a85dd17f..f777906d54e 100644 --- a/tests/e2e/helpers/orca-restart.ts +++ b/tests/e2e/helpers/orca-restart.ts @@ -245,11 +245,7 @@ export async function attachRepoAndOpenTerminal(page: Page, repoPath: string): P return false } // Show the seeded Git worktree external to Orca’s managed workspace root. - await store.getState().updateRepo(repo.id, { - externalWorktreeVisibility: 'show', - // Restart specs model an existing repo, which keeps pre-trust. - agentWorkspaceTrust: 'pre-trust' - }) + await store.getState().updateRepo(repo.id, { externalWorktreeVisibility: 'show' }) return true }, repoId) ), From fd20d34362d66dfa96fee60c09769a5a415c75f9 Mon Sep 17 00:00:00 2001 From: Brennan Benson <79079362+brennanb2025@users.noreply.github.com> Date: Tue, 10 Nov 2026 16:38:44 -0800 Subject: [PATCH 05/21] Save the issue title on GitHub issue links made without one (#40853) * fix(github): title GitHub issue links that were saved without one A workspace linked to a GitHub issue from the CLI (`--issue N`), a phone, the paired web client or a typed number kept only the number, so search could match its issue title only while a client happened to hold that issue. The host that owns the workspace now reads each newly linked untitled issue once, after the write returns, the way its card would, and saves the title on the link. It keeps the current review selection and never resurrects a removed link. * fix(github): bound and publish issue-title fills like other metadata writes Skip writes that leave links untouched before normalizing, check a queued link is still wanted before reading it, publish filled titles once per repo per 2 s window, refresh the resolved rows remote clients read, never wake a stopped WSL distro for a background read, keep a selected review's PR suppression, and keep a failing write observer from failing a saved write. * fix(github): keep link-title fills off unrelated writes and out of per-title refreshes Ignore linkedWorkItem in the unchanged-links check (it is rebuilt on every write and always titled), publish filled titles at once on a quiet host and at most once per repo per 30 s during a burst on a timer, and only lift a PR suppression when the selected PR changes, so restating the current PR (as a title fill does) keeps a hidden branch PR hidden. * perf(github): publish filled link titles as addressed workspace metadata Instead of a whole-repo worktreesChanged refresh per burst window, the host publishes one addressed workspaceMetadata set record {hostId, epoch, seq, id, row} for the workspace whose link was titled: to its desktop window over IPC and to paired clients as a runtime client event. Clients patch that one row (Cmd-J reads it) and drop replayed records; older clients ignore the new event and see the title on their next listing. * fix(github): agreed workspaceMetadata record shape, per-id seqs and late-listing refill Records name their kind and domain, clients keep the last applied seq per workspace for each producer epoch (dropped when the epoch changes), and a listing that read metadata before the title write cannot drop the title: recorded titles refill rows that still hold the same untitled link for 60 s, never restoring a removed link. The IPC channel is marked as interim transport. * fix(github): title-only, host-addressed application of workspaceMetadata records Clients fill the record's issue titles into the named row's still-untitled links and never take its links over the row's current ones, so a record that arrives after an unlink cannot restore the issue. A paired host's records reach the row of their own execution host (an SSH twin is a different row), and go through the bridge's connection-generation check like other host events. * test(github): type the metadata-record test helper as AppState * fix(github): keep recorded titles until expiry, refill only changed repos, retire with the pairing A record that beats the listing adding its link keeps its titles for the full window; the refill runs only for repos whose listing changed, indexes that repo's rows once and skips rows already checked; per-workspace seqs live in the expiring entries; a re-pair or reconnect of the producing environment retires its titles. * fix(github): read the listing through the workspace record selector * fix(github): keep link titles until their producer changes; fall back to a listing on eviction Recorded issue titles no longer expire on a timer (a slow listing could outlast it): they live until the producer's epoch, pairing or connection changes, refilling only present untitled links. When more than the cap are in flight the oldest is evicted and its repo is listed afresh, which carries the saved title. The refill runs in updater form so it never merges a stale store snapshot. * fix(github): sequenced listings replace the retained-title refill Per the re-plan ruling: the store keeps one workspaceMetadata cursor per process; a title record takes the next seq in the same turn as its write, and each listing records the cursor when it reads metadata. Clients note each record's seq per producer and repo; a listing at or past it ends the obligation, an older one keeps titles the rows already show and lists the repo afresh (a new invocation, never joining the stale one), a record whose row or link isn't listed yet asks for a fresh listing, and a seq gap lists every repo of that producer. The retained-title table, its watcher, TTL and cap are gone. * test(github): stamp capture, gap, restart and recovery coverage for workspace metadata listings Also retire an environment's earlier coverage bookkeeping when it is re-paired or reconnects. * fix(space): keep issue-title coverage per repo and host until a newer listing lands Coverage now keeps each projection's newest record seq after its covering listing settles, so an older listing that finishes later still keeps shown titles. Projections are keyed by repo and host, paired producers keep their stream cursor, retired pairings and unlisted projections are dropped, every listing served by this host carries its metadata stamp, and a record with any unlisted titled link asks for a listing. * fix(space): re-list after any stale listing and drop retired metadata producers A stale listing now always asks for one fresh listing, so a link it dropped comes back and a failed scan can no longer end the wait. Removed environments drop their producer state, the producer's repo list is built only on a gap or a first-seen projection, and recovery listings skip the lineage refresh. * perf(space): filter current rows only for listings that keep titles * fix(space): forget removed repos on a metadata gap; keep titles for another producer run's listing * fix(palette): search a linked issue by its saved title before its card renders The issue's search entry took its title only from the issue cache, which fills when the card renders, and it shadowed the saved link's titled entry under the same id. It now falls back to the link's saved title. * perf(palette): read a row's links once per palette document * fix(palette): give each saved link its own search unit Saved-link evidence was meant to carry an `attachment:` id but the override set an unused field, so a link shared the `issue:N` unit id with the legacy issue field and with same-number links from other repos, and only the first (often untitled) survived. The id now applies to the unit and its fields, so every saved link's title is searchable. Replaces the cache-title fallback. * fix(space): keep the title queue draining past a failed link; encode environment ids in coverage keys --------- --- .../github-issue-link-title-fill.test.ts | 114 +++++++++ .../github/github-issue-link-title-fill.ts | 166 +++++++++++++ .../github/github-issue-link-title-wiring.ts | 103 ++++++++ ...d-provider-listing-catalog-version.test.ts | 27 +- .../listing/detected-provider-listing.ts | 26 +- src/main/orcad/orcad-entry.ts | 2 + .../loading-store/store-runtime-state.ts | 6 +- .../worktree-identity-metadata.ts | 11 +- .../worktree-meta-write-normalization.ts | 2 +- .../loading-store/write-scheduling.ts | 32 +++ .../events/runtime-notifier-contract.ts | 2 + .../construction/runtime-store-contract.ts | 1 + .../catalog/runtime-listing-metadata.ts | 28 +++ .../runtime-managed-worktree-queries.ts | 16 +- .../runtime-workspace-change-events.ts | 7 + .../startup/main-process-runtime-service.ts | 4 + src/main/window/runtime-window-lifecycle.ts | 2 + src/preload/api/worktree-api.ts | 2 + src/preload/api/worktrees-bridge.ts | 7 + .../ipc-events/project-catalog-ipc-bridge.ts | 5 + .../ipc-events/runtime-client-ipc-bridge.ts | 5 +- .../worktree-metadata-set-events.ts | 98 ++++++++ .../worktrees/worktree-palette-document.ts | 14 +- .../worktrees/worktree-palette-search.test.ts | 30 +++ .../subscriptions/runtime-client-events.ts | 1 + .../detected-worktree-refresh-leases.ts | 23 +- .../listing/detected-worktree-refresh.ts | 25 +- .../worktrees/listing/fetch-all-worktrees.ts | 7 +- .../worktrees/listing/fetch-worktrees.ts | 10 +- .../workspace-metadata-coverage.test.ts | 76 ++++++ .../listing/workspace-metadata-coverage.ts | 235 ++++++++++++++++++ .../listing/workspace-metadata-recovery.ts | 40 +++ .../listing/worktree-fetch-options.ts | 2 + .../worktree-metadata-set-patch.test.ts | 124 +++++++++ .../listing/worktree-metadata-set-patch.ts | 83 +++++++ .../worktrees/listing/worktree-slice-types.ts | 2 + .../metadata/update-worktree-meta.ts | 3 +- .../src/web/preload-api/web-worktrees-api.ts | 1 + src/shared/runtime-client-events.ts | 3 + src/shared/worktree/github-pr-suppression.ts | 8 +- src/shared/worktree/types.ts | 2 + src/shared/worktree/workspace-metadata-set.ts | 84 +++++++ 42 files changed, 1392 insertions(+), 47 deletions(-) create mode 100644 src/main/github/github-issue-link-title-fill.test.ts create mode 100644 src/main/github/github-issue-link-title-fill.ts create mode 100644 src/main/github/github-issue-link-title-wiring.ts create mode 100644 src/main/runtime/workspaces/catalog/runtime-listing-metadata.ts create mode 100644 src/renderer/src/hooks/ipc-events/worktree-metadata-set-events.ts create mode 100644 src/renderer/src/store/slices/worktrees/listing/workspace-metadata-coverage.test.ts create mode 100644 src/renderer/src/store/slices/worktrees/listing/workspace-metadata-coverage.ts create mode 100644 src/renderer/src/store/slices/worktrees/listing/workspace-metadata-recovery.ts create mode 100644 src/renderer/src/store/slices/worktrees/listing/worktree-metadata-set-patch.test.ts create mode 100644 src/renderer/src/store/slices/worktrees/listing/worktree-metadata-set-patch.ts create mode 100644 src/shared/worktree/workspace-metadata-set.ts diff --git a/src/main/github/github-issue-link-title-fill.test.ts b/src/main/github/github-issue-link-title-fill.test.ts new file mode 100644 index 00000000000..f70b9c2520e --- /dev/null +++ b/src/main/github/github-issue-link-title-fill.test.ts @@ -0,0 +1,114 @@ +import { describe, it, expect, vi, beforeEach, afterEach } from 'vitest' +import { mkdtempSync, rmSync } from 'node:fs' +import { join } from 'node:path' +import { tmpdir } from 'node:os' +import { + closeTestStores, + createStore, + makeRepo, + testState, + writeDataFile +} from '../persistence/persistence-test-harness' +import { getWorkspaceAttachments } from '../../shared/workspace-attachments/workspace-attachments' +import { hostWorktreeTarget } from '../../shared/worktree/host-qualified-identity' +import { installGitHubIssueLinkTitleFill } from './github-issue-link-title-fill' + +vi.mock('electron', () => ({ + app: { getPath: () => testState.dir }, + safeStorage: { isEncryptionAvailable: () => false } +})) +vi.mock('../telemetry/client', () => ({ track: vi.fn() })) +vi.mock('../telemetry/cohort-classifier', () => ({ getCohortAtEmit: vi.fn() })) + +const target = hostWorktreeTarget({ id: 'repo::/work', hostId: 'local' }) + +async function setup(readIssueTitle: () => Promise) { + writeDataFile({ repos: [makeRepo({ id: 'repo', path: '/work' })] }) + const store = await createStore() + const pending: (() => void)[] = [] + const publish = vi.fn() + installGitHubIssueLinkTitleFill({ + store, + readIssueTitle, + publish, + schedule: (run) => pending.push(run) + }) + const settle = async (): Promise => { + for (const run of pending.splice(0)) { + run() + } + await vi.waitFor(() => expect(readIssueTitle).toHaveBeenCalled()) + await new Promise((resolve) => setImmediate(resolve)) + } + return { store, settle, publish } +} + +describe('GitHub issue link title fill', () => { + beforeEach(() => { + testState.dir = mkdtempSync(join(tmpdir(), 'orca-issue-link-title-')) + }) + afterEach(async () => { + await closeTestStores() + rmSync(testState.dir, { recursive: true, force: true }) + }) + + it('titles a bare linked issue without re-selecting a review the user had cleared', async () => { + const { store, settle, publish } = await setup(vi.fn(async () => 'Repair authentication')) + // A kept merge request with no review selected, as after the user clears the selection. + store.setWorktreeMeta(target, { + linkedItems: [{ provider: 'gitlab', type: 'mr', number: 5 }], + linkedPR: null, + linkedGitLabMR: null, + linkedBitbucketPR: null, + linkedAzureDevOpsPR: null, + linkedGiteaPR: null + }) + store.setWorktreeMeta(target, { linkedIssue: 42 }) + expect(store.getWorktreeMetaForHost('repo::/work', 'local')?.linkedGitLabMR ?? null).toBeNull() + await settle() + + const meta = store.getWorktreeMetaForHost('repo::/work', 'local') + expect(getWorkspaceAttachments(meta)).toContainEqual( + expect.objectContaining({ + provider: 'github', + type: 'issue', + number: 42, + title: 'Repair authentication' + }) + ) + expect(meta?.linkedIssue).toBe(42) + expect(meta?.linkedGitLabMR ?? null).toBeNull() + expect(publish).toHaveBeenCalledWith( + expect.objectContaining({ worktreeId: 'repo::/work', hostId: 'local' }) + ) + }) + + it('keeps a hidden branch PR hidden when titling a link beside the selected PR', async () => { + const { store, settle } = await setup(vi.fn(async () => 'Repair authentication')) + store.setWorktreeMeta(target, { linkedPR: 7 }) + store.setWorktreeMeta(target, { suppressedGitHubPR: 7 }) + store.setWorktreeMeta(target, { linkedIssue: 42 }) + await settle() + + const meta = store.getWorktreeMetaForHost('repo::/work', 'local') + expect(getWorkspaceAttachments(meta)).toContainEqual( + expect.objectContaining({ type: 'issue', number: 42, title: 'Repair authentication' }) + ) + expect(meta?.suppressedGitHubPR).toBe(7) + }) + + it('does not bring back a link removed while its title was being read', async () => { + let finish: (title: string) => void = () => {} + const read = vi.fn(() => new Promise((resolve) => (finish = resolve))) + const { store, settle } = await setup(read) + store.setWorktreeMeta(target, { linkedIssue: 42 }) + const settled = settle() + await vi.waitFor(() => expect(read).toHaveBeenCalled()) + store.setWorktreeMeta(target, { linkedIssue: null }) + finish('Repair authentication') + await settled + + const links = getWorkspaceAttachments(store.getWorktreeMetaForHost('repo::/work', 'local')) + expect(links.filter((item) => item.provider === 'github' && item.type === 'issue')).toEqual([]) + }) +}) diff --git a/src/main/github/github-issue-link-title-fill.ts b/src/main/github/github-issue-link-title-fill.ts new file mode 100644 index 00000000000..bea7dcc0f3e --- /dev/null +++ b/src/main/github/github-issue-link-title-fill.ts @@ -0,0 +1,166 @@ +import type { ExecutionHostId } from '../../shared/execution-host' +import type { Repo } from '../../shared/repo-types' +import type { WorktreeMeta } from '../../shared/worktree/meta-types' +import type { WorkspaceAttachment } from '../../shared/worktree/types' +import type { WorkspaceAttachmentMutation } from '../../shared/workspace-attachments/workspace-attachment-mutation' +import { getWorkspaceAttachments } from '../../shared/workspace-attachments/workspace-attachments' +import { WORKSPACE_ATTACHMENT_NUMBER_SLOTS } from '../../shared/workspace-attachments/workspace-attachment-legacy' +import { getWorkspaceReferenceIdentity } from '../../shared/workspace-attachments/workspace-reference-identity' +import { + hostWorktreeTarget, + type HostWorktreeTarget +} from '../../shared/worktree/host-qualified-identity' +import { getRepoIdFromWorktreeId } from '../../shared/worktree/id' +import type { WorktreeMetaWriteListener } from '../persistence/loading-store/write-scheduling' +import { findRepoRowForHostScopedWrite } from '../persistence/tracking-repos/repo-update-operations' + +export type GitHubIssueLinkTitleFillStore = { + onWorktreeMetaWritten: (listener: WorktreeMetaWriteListener) => () => void + getWorktreeMetaForHost: (worktreeId: string, hostId: ExecutionHostId) => WorktreeMeta | undefined + setWorktreeMeta: ( + target: HostWorktreeTarget, + meta: Partial & WorkspaceAttachmentMutation + ) => WorktreeMeta + getRepos: () => Repo[] +} + +type Job = { + worktreeId: string + hostId: ExecutionHostId + instanceId: string | undefined + identity: string + link: WorkspaceAttachment +} + +function untitledGitHubIssues(meta: WorktreeMeta | undefined): WorkspaceAttachment[] { + return getWorkspaceAttachments(meta).filter( + (item) => item.provider === 'github' && item.type === 'issue' && !item.title?.trim() + ) +} + +/** + * Links made without a title (CLI `--issue N`, phones, typed numbers) get the issue's title once, on the + * host that owns the workspace, so search can match it without a client re-reading every linked issue. + */ +export function installGitHubIssueLinkTitleFill(deps: { + store: GitHubIssueLinkTitleFillStore + readIssueTitle: (repo: Repo, link: WorkspaceAttachment) => Promise + // Addressed to the one workspace whose links changed; never a repo-wide refresh. + publish: (change: { worktreeId: string; hostId: ExecutionHostId; meta: WorktreeMeta }) => void + schedule?: (run: () => void) => void +}): () => void { + const schedule = deps.schedule ?? ((run) => setImmediate(run)) + const queue = new Map() + let draining = false + let disposed = false + + // The job's link as it is now; null once removed, retitled or its workspace re-created. + const currentTarget = (job: Job) => { + const meta = deps.store.getWorktreeMetaForHost(job.worktreeId, job.hostId) + if (!meta || meta.instanceId !== job.instanceId) { + return null + } + const items = getWorkspaceAttachments(meta) + const index = items.findIndex( + (item) => getWorkspaceReferenceIdentity(item) === job.identity && !item.title?.trim() + ) + return index === -1 ? null : { meta, items, index } + } + + const apply = (job: Job, title: string): void => { + const target = currentTarget(job) + if (!target) { + return + } + const { meta, items, index } = target + const reviewSlots = WORKSPACE_ATTACHMENT_NUMBER_SLOTS.filter(([, , type]) => type !== 'issue') + // A cleared review selection must be restated or the merge re-picks one; a set one it keeps. + const keepClearedSelection = reviewSlots.every(([slot]) => (meta[slot] ?? null) === null) + ? Object.fromEntries(reviewSlots.map(([slot]) => [slot, null])) + : {} + const updated = deps.store.setWorktreeMeta( + hostWorktreeTarget({ id: job.worktreeId, hostId: job.hostId }), + { + ...keepClearedSelection, + linkedItemsBase: items, + linkedItems: items.map((item, i) => (i === index ? { ...item, title } : item)) + } + ) + deps.publish({ worktreeId: job.worktreeId, hostId: job.hostId, meta: updated }) + } + + const drain = async (): Promise => { + if (draining) { + return + } + draining = true + try { + for (const [key, job] of queue) { + queue.delete(key) + if (disposed) { + return + } + try { + const repo = currentTarget(job) + ? findRepoRowForHostScopedWrite( + deps.store.getRepos(), + getRepoIdFromWorktreeId(job.worktreeId), + job.hostId + ) + : undefined + const title = repo ? await deps.readIssueTitle(repo, job.link).catch(() => null) : null + if (title?.trim() && !disposed) { + apply(job, title.trim()) + } + } catch { + // One link whose target changed underneath it (e.g. an alias turned ambiguous) is skipped; + // the rest of the queue still drains. + } + } + } finally { + draining = false + } + } + + const unsubscribe = deps.store.onWorktreeMetaWritten(({ worktreeId, hostId, previous, next }) => { + // Most writes (stamps, flags, comments) leave links untouched; skip them without normalizing. + // linkedWorkItem is re-created on every write and always titled, so it is not compared. + if ( + previous && + previous.instanceId === next.instanceId && + previous.linkedItems === next.linkedItems && + previous.linkedIssue === next.linkedIssue + ) { + return + } + const known = new Set( + previous?.instanceId === next.instanceId + ? untitledGitHubIssues(previous).map(getWorkspaceReferenceIdentity) + : [] + ) + let added = false + for (const link of untitledGitHubIssues(next)) { + const identity = getWorkspaceReferenceIdentity(link) + if (known.has(identity)) { + continue + } + queue.set(JSON.stringify([hostId, worktreeId, next.instanceId, identity]), { + worktreeId, + hostId, + instanceId: next.instanceId, + identity, + link + }) + added = true + } + if (added) { + // Run after the caller's write returns; nothing the user asked for waits on GitHub. + schedule(() => void drain()) + } + }) + return () => { + disposed = true + queue.clear() + unsubscribe() + } +} diff --git a/src/main/github/github-issue-link-title-wiring.ts b/src/main/github/github-issue-link-title-wiring.ts new file mode 100644 index 00000000000..0c22857efcf --- /dev/null +++ b/src/main/github/github-issue-link-title-wiring.ts @@ -0,0 +1,103 @@ +import type { Repo } from '../../shared/repo-types' +import { composeWorktreeHostIdentity } from '../../shared/worktree/host-qualified-identity' +import { getRepoIdFromWorktreeId } from '../../shared/worktree/id' +import { + workspaceMetadataSetRow, + type WorkspaceMetadataSet +} from '../../shared/worktree/workspace-metadata-set' +import type { WorkspaceAttachment } from '../../shared/worktree/types' +import { getRepoRoutingConnectionId } from '../../shared/execution-host' +import type { Store } from '../persistence/loading-store/store' +import { getLocalProjectGhExecOptions } from '../workspaces/execution-host/project-runtime-git-options' +import { getIssue } from './issues' +import { getWorkItemByOwnerRepo } from './client/fetch/get-work-item' +import { installGitHubIssueLinkTitleFill } from './github-issue-link-title-fill' +import { readGitHubRepoUnlessWslStopped } from '../ipc/source-control/github-repo-routing' + +const GITHUB_ISSUE_PATH = /^\/([^/]+)\/([^/]+)\/issues\/(\d+)$/ + +function sameIssueUrl(left: string, right: string): boolean { + try { + const a = new URL(left) + const b = new URL(right) + return ( + a.hostname.toLowerCase() === b.hostname.toLowerCase() && + a.pathname.replace(/\/+$/, '').toLowerCase() === b.pathname.replace(/\/+$/, '').toLowerCase() + ) + } catch { + return false + } +} + +/** Reads a link's issue the way its display does: a URL names its repo; a bare number uses the card's lookup. */ +async function readGitHubIssueLinkTitle( + store: Store, + repo: Repo, + link: WorkspaceAttachment +): Promise { + const connectionId = getRepoRoutingConnectionId(repo) + const options = getLocalProjectGhExecOptions(store, repo) + if (link.url) { + const url = new URL(link.url) + const route = GITHUB_ISSUE_PATH.exec(url.pathname) + if (!route || Number(route[3]) !== link.number) { + return null + } + const item = await getWorkItemByOwnerRepo( + repo.path, + { owner: route[1], repo: route[2], host: url.hostname }, + link.number, + 'issue', + connectionId, + options, + { preview: true } + ) + return item?.type === 'issue' && item.number === link.number && sameIssueUrl(item.url, link.url) + ? item.title + : null + } + const issue = await getIssue(repo.path, link.number, connectionId, options) + // The issues endpoint also answers PR numbers; a PR's title must not label an issue link. + return issue && issue.number === link.number && !/\/pull\/\d+$/.test(issue.url) + ? issue.title + : null +} + +export function wireGitHubIssueLinkTitleFill( + store: Store, + runtime: { + invalidateResolvedWorktreeCache: () => void + publishWorkspaceMetadataSet: (set: WorkspaceMetadataSet) => void + } +): () => void { + return installGitHubIssueLinkTitleFill({ + store, + // Background reads never start a stopped WSL distro; the link just stays untitled. + readIssueTitle: (repo, link) => + readGitHubRepoUnlessWslStopped( + store, + repo, + () => readGitHubIssueLinkTitle(store, repo, link), + { + wake: false + } + ), + publish: ({ worktreeId, hostId, meta }) => { + // Resolved rows embed metadata; refresh them so the next listing read agrees. + runtime.invalidateResolvedWorktreeCache() + // Same turn as the title write, so a listing either read it or carries an older seq. + const { epoch, seq } = store.advanceWorkspaceMetadataCursor() + runtime.publishWorkspaceMetadataSet({ + kind: 'set', + domain: 'workspaceMetadata', + hostId, + epoch, + seq, + id: composeWorktreeHostIdentity(hostId, worktreeId), + repoId: getRepoIdFromWorktreeId(worktreeId), + worktreeId, + row: workspaceMetadataSetRow(meta) + }) + } + }) +} diff --git a/src/main/ipc/worktrees/listing/detected-provider-listing-catalog-version.test.ts b/src/main/ipc/worktrees/listing/detected-provider-listing-catalog-version.test.ts index 3ea51313aa2..c9d96275b1d 100644 --- a/src/main/ipc/worktrees/listing/detected-provider-listing-catalog-version.test.ts +++ b/src/main/ipc/worktrees/listing/detected-provider-listing-catalog-version.test.ts @@ -41,7 +41,7 @@ const localRepo: Repo = { addedAt: 0 } -function createStore(repo: Repo) { +function createStore(repo: Repo, cursor?: { epoch: string; seq: number }) { // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: the listing reads only these store methods; the pruning that would need more is mocked above. return { getRepo: () => repo, @@ -55,7 +55,8 @@ function createStore(repo: Repo) { getAllWorktreeLineage: () => ({}), getAllWorkspaceLineage: () => ({}), removeWorktreeLineage: vi.fn(), - captureNativeLocalWorktreeMetadataScanExpectation: () => undefined + captureNativeLocalWorktreeMetadataScanExpectation: () => undefined, + ...(cursor ? { getWorkspaceMetadataCursor: () => ({ ...cursor }) } : {}) } as never } @@ -106,4 +107,26 @@ describe('the catalog version a listing carries', () => { first && 'catalogVersion' in first ? first.catalogVersion : undefined ) }) + + it('carries the metadata cursor it read with its metadata, not one advanced during the scan', async () => { + const cursor = { epoch: 'e1', seq: 4 } + // A title write lands while the scan is still running. + listRepoWorktreesMock.mockImplementation(async () => { + cursor.seq = 5 + return [worktreeAt(localRepo, localRepo.path)] + }) + const store = createStore(localRepo, cursor) + + const during = await listDetectedWorktreesForCapturedRepo(store, localRepo, () => true) + const after = await listDetectedWorktreesForCapturedRepo(store, localRepo, () => true) + + expect(during && 'workspaceMetadata' in during ? during.workspaceMetadata : null).toEqual({ + epoch: 'e1', + seq: 4 + }) + expect(after && 'workspaceMetadata' in after ? after.workspaceMetadata : null).toEqual({ + epoch: 'e1', + seq: 5 + }) + }) }) diff --git a/src/main/ipc/worktrees/listing/detected-provider-listing.ts b/src/main/ipc/worktrees/listing/detected-provider-listing.ts index 3344d22e1a6..7aa21b9e99a 100644 --- a/src/main/ipc/worktrees/listing/detected-provider-listing.ts +++ b/src/main/ipc/worktrees/listing/detected-provider-listing.ts @@ -46,6 +46,13 @@ import { wslDistroNotRunningForScanFailure } from './wsl-stopped-worktree-listing' +// Partial stores (tests, embedders) may not publish workspace metadata; their listings carry no stamp. +function readWorkspaceMetadataCursor(store: { + getWorkspaceMetadataCursor?: () => { epoch: string; seq: number } +}): { epoch: string; seq: number } | undefined { + return store.getWorkspaceMetadataCursor?.() +} + export async function listDetectedWorktreesForCapturedRepo( store: Store, repo: Repo, @@ -59,6 +66,8 @@ export async function listDetectedWorktreesForCapturedRepo( ? ({ providerAbortStatus: providerAbort.status() } as const) : undefined const allMeta = isFolderRepo(repo) ? undefined : readAllWorktreeMetaForRepo(store, repo) + // Read with the metadata, so clients can tell which published metadata records this listing includes. + const workspaceMetadata = readWorkspaceMetadataCursor(store) try { // Folder rows are read synchronously, leaving no mutation window that needs a re-scan. @@ -75,7 +84,8 @@ export async function listDetectedWorktreesForCapturedRepo( authoritative: false, source: 'metadata-fallback', worktrees: [], - catalogVersion: getLocalWorktreeCatalogVersion(repo.id) + catalogVersion: getLocalWorktreeCatalogVersion(repo.id), + ...(workspaceMetadata ? { workspaceMetadata } : {}) } } return { @@ -83,6 +93,7 @@ export async function listDetectedWorktreesForCapturedRepo( authoritative: true, source: 'git', catalogVersion: getLocalWorktreeCatalogVersion(repo.id), + ...(workspaceMetadata ? { workspaceMetadata } : {}), worktrees: projectWorktreeLineageWithHostPeers( store, repo, @@ -108,7 +119,8 @@ export async function listDetectedWorktreesForCapturedRepo( authoritative: false, source: 'metadata-fallback', worktrees: buildDisconnectedDetectedWorktrees(store, repo, worktrees), - catalogVersion: getLocalWorktreeCatalogVersion(repo.id) + catalogVersion: getLocalWorktreeCatalogVersion(repo.id), + ...(workspaceMetadata ? { workspaceMetadata } : {}) } } const pendingAtScan = snapshotPendingWorktreeRemovals() @@ -140,7 +152,8 @@ export async function listDetectedWorktreesForCapturedRepo( authoritative: false, source: 'metadata-fallback', worktrees: [], - catalogVersion: localWorktreeCatalogVersionAt(scan.generation) + catalogVersion: localWorktreeCatalogVersionAt(scan.generation), + ...(workspaceMetadata ? { workspaceMetadata } : {}) } } if (freshScan) { @@ -166,6 +179,7 @@ export async function listDetectedWorktreesForCapturedRepo( authoritative: true, source: 'git', catalogVersion: localWorktreeCatalogVersionAt(scan.generation), + ...(workspaceMetadata ? { workspaceMetadata } : {}), worktrees: projectPendingWorktreeRemovals( buildDetectedGitWorktrees(store, repo, localRows, allMeta), (worktree) => worktree.id, @@ -196,7 +210,8 @@ export async function listDetectedWorktreesForCapturedRepo( unavailableReason, failureKind, wslDistroStopped, - catalogVersion: getLocalWorktreeCatalogVersion(repo.id) + catalogVersion: getLocalWorktreeCatalogVersion(repo.id), + ...(workspaceMetadata ? { workspaceMetadata } : {}) } } warnOnce( @@ -212,7 +227,8 @@ export async function listDetectedWorktreesForCapturedRepo( worktrees: [], unavailableReason, failureKind, - catalogVersion: getLocalWorktreeCatalogVersion(repo.id) + catalogVersion: getLocalWorktreeCatalogVersion(repo.id), + ...(workspaceMetadata ? { workspaceMetadata } : {}) } } } diff --git a/src/main/orcad/orcad-entry.ts b/src/main/orcad/orcad-entry.ts index 0f463081e63..da51c916fe2 100644 --- a/src/main/orcad/orcad-entry.ts +++ b/src/main/orcad/orcad-entry.ts @@ -226,6 +226,8 @@ async function startOrcadRuntime( paneGridsForShutdown = runtime accountLaunch?.attachToRuntime(runtime) + const { wireGitHubIssueLinkTitleFill } = await import('../github/github-issue-link-title-wiring') + registerCleanup(wireGitHubIssueLinkTitleFill(profileStore, runtime)) const { installOrcadEmulatorBridge } = await import('./orcad-emulator-bridge') installOrcadEmulatorBridge(runtime, registerCleanup) diff --git a/src/main/persistence/loading-store/store-runtime-state.ts b/src/main/persistence/loading-store/store-runtime-state.ts index 951187da47b..d089fa159a4 100644 --- a/src/main/persistence/loading-store/store-runtime-state.ts +++ b/src/main/persistence/loading-store/store-runtime-state.ts @@ -1,3 +1,4 @@ +import { randomUUID } from 'node:crypto' import { removeStaleDurableWriteTempFiles } from '../../filesystem/durable-file-write' import type { GlobalSettings } from '../../../shared/global-settings-types' import type { PersistedState } from '../../../shared/persisted-state-types' @@ -6,7 +7,7 @@ import type { WorkspaceSessionWriteListener } from './workspace-session-write-listeners' import type { ExecutionHostId } from '../../../shared/execution-host' -import type { StateWriteListener } from './write-scheduling' +import type { StateWriteListener, WorktreeMetaWriteListener } from './write-scheduling' import type { PartitionEditorViews } from './layout-editor-views' import type { ActiveViewPreference } from '../preferences/active-view-preference' import { @@ -114,6 +115,9 @@ export class StoreRuntimeState { /** In-memory editor views by partition; layout-owner builds only. */ readonly layoutEditorViews = new Map() readonly stateWriteListeners = new Set() + readonly worktreeMetaWriteListeners = new Set() + // This store's workspaceMetadata position: a new epoch per process, one seq per published record. + readonly workspaceMetadataCursor = { epoch: randomUUID(), seq: 0 } projectHostOperations: ProjectHostPersistenceOperations | null = null projectGroupOperations: ProjectGroupPersistenceOperations | null = null folderWorkspaceOperations: FolderWorkspacePersistenceOperations | null = null diff --git a/src/main/persistence/loading-store/worktree-identity-metadata.ts b/src/main/persistence/loading-store/worktree-identity-metadata.ts index 8a9a78369b8..5b2c74cf061 100644 --- a/src/main/persistence/loading-store/worktree-identity-metadata.ts +++ b/src/main/persistence/loading-store/worktree-identity-metadata.ts @@ -19,7 +19,8 @@ import { worktreeMetaHostFor } from '../../../shared/worktree/local-only-worktree-meta-ids' -type MetadataRuntime = Pick +type MetadataRuntime = Pick & + Partial> type WorktreeMetaState = Pick< PersistedState, 'worktreeMeta' | 'worktreeMetaByIdentity' | 'worktreeIdentityAliases' @@ -265,5 +266,13 @@ export function setWorktreeMetaForHost( state.worktreeMeta[worktreeId] = updated } scheduleSave(scheduling, WORKTREE_METADATA_DOMAINS) + for (const listener of runtime.worktreeMetaWriteListeners ?? []) { + try { + listener({ worktreeId, hostId: executionHostId, previous: existing, next: updated }) + } catch (error) { + // The row is already written; an observer's failure must not report the write as failed. + console.error('[worktree-meta] write listener failed:', error) + } + } return updated } diff --git a/src/main/persistence/loading-store/worktree-meta-write-normalization.ts b/src/main/persistence/loading-store/worktree-meta-write-normalization.ts index 638623f5202..e98a0e2e7f7 100644 --- a/src/main/persistence/loading-store/worktree-meta-write-normalization.ts +++ b/src/main/persistence/loading-store/worktree-meta-write-normalization.ts @@ -49,7 +49,7 @@ export function mergeWorktreeMetaForWrite( const normalizedUpdates = normalizeWorkspaceAttachmentUpdate(existing, metaUpdates) const updated = { ...(existing ?? createDefaultWorktreeMeta()), - ...normalizeGitHubPRSuppressionUpdate(normalizedUpdates), + ...normalizeGitHubPRSuppressionUpdate(normalizedUpdates, existing?.linkedPR), ...identity } updated.linkedWorkItem = normalizeWorkspaceLinkedItem(updated.linkedWorkItem) diff --git a/src/main/persistence/loading-store/write-scheduling.ts b/src/main/persistence/loading-store/write-scheduling.ts index a5bd9ccd810..7397530f9ac 100644 --- a/src/main/persistence/loading-store/write-scheduling.ts +++ b/src/main/persistence/loading-store/write-scheduling.ts @@ -9,9 +9,17 @@ import { } from './workspace-session-write-listeners' import { ptyBindingGenerationOf } from './pty-binding-generations' import type { PartitionEditorViews } from './layout-editor-views' +import type { WorktreeMeta } from '../../../shared/worktree/meta-types' /** Undefined means any domain may have changed. */ export type StateWriteListener = (dirtyDomains: readonly string[] | undefined) => void +/** One host's worktree metadata row before and after a write. */ +export type WorktreeMetaWriteListener = (write: { + worktreeId: string + hostId: ExecutionHostId + previous: WorktreeMeta | undefined + next: WorktreeMeta +}) => void const SAVE_DEBOUNCE_MS = 1_000 const SAVE_MAX_WAIT_MS = 5_000 @@ -31,6 +39,8 @@ type WriteSchedulingOperationsRuntime = Pick< | 'workspaceSessionWriteListeners' | 'workspaceSessionOperationListeners' | 'stateWriteListeners' + | 'worktreeMetaWriteListeners' + | 'workspaceMetadataCursor' | 'ptyBindingGenerations' | 'layoutEditorViews' | 'state' @@ -88,6 +98,28 @@ export class WriteSchedulingOperations { runtime.stateWriteListeners.delete(listener) } } + + /** The workspaceMetadata position a listing records when it reads metadata. */ + getWorkspaceMetadataCursor(): { epoch: string; seq: number } { + const { epoch, seq } = this[writeSchedulingOperationsContext].runtime.workspaceMetadataCursor + return { epoch, seq } + } + + /** Advances the position for one published record; call in the same turn as its metadata write. */ + advanceWorkspaceMetadataCursor(): { epoch: string; seq: number } { + const cursor = this[writeSchedulingOperationsContext].runtime.workspaceMetadataCursor + cursor.seq += 1 + return { epoch: cursor.epoch, seq: cursor.seq } + } + + /** Observe each host-qualified worktree metadata write; callbacks must not throw, block or write. */ + onWorktreeMetaWritten(listener: WorktreeMetaWriteListener): () => void { + const { runtime } = this[writeSchedulingOperationsContext] + runtime.worktreeMetaWriteListeners.add(listener) + return () => { + runtime.worktreeMetaWriteListeners.delete(listener) + } + } } export function scheduleSave( diff --git a/src/main/runtime/client/events/runtime-notifier-contract.ts b/src/main/runtime/client/events/runtime-notifier-contract.ts index 699cea69893..0715c88c511 100644 --- a/src/main/runtime/client/events/runtime-notifier-contract.ts +++ b/src/main/runtime/client/events/runtime-notifier-contract.ts @@ -1,3 +1,4 @@ +import type { WorkspaceMetadataSet } from '../../../../shared/worktree/workspace-metadata-set' import type { RuntimeEditorTabReveal, RuntimeFileOpenPosition @@ -49,6 +50,7 @@ export type RuntimeNotifier = { // Why: false for metadata-only changes (sort order) that leave Git scan results valid. options?: { invalidateGitScans?: boolean } ): void + workspaceMetadataSet?(set: WorkspaceMetadataSet): void worktreeBaseStatus?(event: WorktreeBaseStatusEvent): void worktreeRemoteBranchConflict?(event: WorktreeRemoteBranchConflictEvent): void reposChanged(): void diff --git a/src/main/runtime/construction/runtime-store-contract.ts b/src/main/runtime/construction/runtime-store-contract.ts index e65b65778b7..4847daa5909 100644 --- a/src/main/runtime/construction/runtime-store-contract.ts +++ b/src/main/runtime/construction/runtime-store-contract.ts @@ -32,6 +32,7 @@ export type RuntimeStore = { removeProjectForHost?: Store['removeProjectForHost'] reorderRepos?: Store['reorderRepos'] getAllWorktreeMeta: Store['getAllWorktreeMeta'] + getWorkspaceMetadataCursor?: Store['getWorkspaceMetadataCursor'] getAllWorktreeMetaForHost?: Store['getAllWorktreeMetaForHost'] captureNativeLocalWorktreeMetadataScanExpectation?: Store['captureNativeLocalWorktreeMetadataScanExpectation'] pruneSessionlessMissingLocalWorktreeMetadataForRepo?: Store['pruneSessionlessMissingLocalWorktreeMetadataForRepo'] diff --git a/src/main/runtime/workspaces/catalog/runtime-listing-metadata.ts b/src/main/runtime/workspaces/catalog/runtime-listing-metadata.ts new file mode 100644 index 00000000000..eff460beecd --- /dev/null +++ b/src/main/runtime/workspaces/catalog/runtime-listing-metadata.ts @@ -0,0 +1,28 @@ +import type { Repo } from '../../../../shared/repo-types' +import type { WorktreeMeta } from '../../../../shared/worktree/meta-types' +import type { WorkspaceMetadataCursor } from '../../../../shared/worktree/workspace-metadata-set' +import type { RuntimeStore } from '../../construction/runtime-store-contract' +import { getRepoOwnedWorktreeMeta } from '../../../workspaces/catalog/worktree-metadata-ownership' + +/** + * The metadata a listing reads, with the store's workspaceMetadata cursor taken in the same turn, so + * clients can tell which published metadata records the listing includes. + */ +export function readRuntimeListingMetadata( + store: RuntimeStore, + repo: Repo +): { + repoOwnerCount: number + workspaceMetadata: WorkspaceMetadataCursor | undefined + // The repo-owned metadata row for a worktree id, from this same read. + ownedMeta: (worktreeId: string) => WorktreeMeta | undefined +} { + const workspaceMetadata = store.getWorkspaceMetadataCursor?.() + const metaById = store.getAllWorktreeMeta() + const repoOwnerCount = store.getRepos().filter((candidate) => candidate.id === repo.id).length + return { + repoOwnerCount, + workspaceMetadata, + ownedMeta: (worktreeId) => getRepoOwnedWorktreeMeta(repo, worktreeId, metaById, repoOwnerCount) + } +} diff --git a/src/main/runtime/workspaces/catalog/runtime-managed-worktree-queries.ts b/src/main/runtime/workspaces/catalog/runtime-managed-worktree-queries.ts index 2aa6864c649..63f4308ea41 100644 --- a/src/main/runtime/workspaces/catalog/runtime-managed-worktree-queries.ts +++ b/src/main/runtime/workspaces/catalog/runtime-managed-worktree-queries.ts @@ -1,3 +1,4 @@ +import { readRuntimeListingMetadata } from './runtime-listing-metadata' import type { RuntimeManagedWorktreeQueryDependencies } from './runtime-managed-worktree-query-dependencies' import { resolveWorktreeHostRouting } from './worktree-launch-host-repo' import { createRepoRowExecutionHostLookup } from '../../../../shared/execution-host/worktree-execution-host-resolution' @@ -137,8 +138,7 @@ export class RuntimeManagedWorktreeQueries { const visibilitySettings = { ...settings, worktreeVisibilityDefaults: visibilityDefaults } if (isFolderRepo(repo)) { const worktrees = listRuntimeFolderWorkspaces(store, repo) - const metaById = store.getAllWorktreeMeta() - const repoOwnerCount = store.getRepos().filter((candidate) => candidate.id === repo.id).length + const listed = readRuntimeListingMetadata(store, repo) const matcher = createWorktreeVisibilitySourceMatcher( [repo.path, ...worktrees.map((worktree) => worktree.path)], resolveCustomWorktreeVisibilitySources(repo, visibilityDefaults), @@ -151,7 +151,7 @@ export class RuntimeManagedWorktreeQueries { matcher, defaultsSupported, visibilitySettings, - getRepoOwnedWorktreeMeta(repo, worktree.id, metaById, repoOwnerCount) ?? null + listed.ownedMeta(worktree.id) ?? null ) ) return { @@ -159,6 +159,7 @@ export class RuntimeManagedWorktreeQueries { authoritative: true, source: 'git', catalogVersion: getLocalWorktreeCatalogVersion(repo.id), + workspaceMetadata: listed.workspaceMetadata, worktrees: projectWorktreeLineageWithHostPeers(store, repo, detected) } } @@ -185,13 +186,11 @@ export class RuntimeManagedWorktreeQueries { resolveConfiguredWorktreeBasePaths(repo) ) const expectedHostId = getRepoExecutionHostId(repo) - const repoOwnerCount = store.getRepos().filter((candidate) => candidate.id === repo.id).length - const metaById = store.getAllWorktreeMeta() + const listed = readRuntimeListingMetadata(store, repo) + const repoOwnerCount = listed.repoOwnerCount const detected = scan.worktrees.map((gitWorktree) => { const id = `${repo.id}::${gitWorktree.path}` - const meta = - readWorktreeMetaForHost(store, id, expectedHostId) ?? - getRepoOwnedWorktreeMeta(repo, id, metaById, repoOwnerCount) + const meta = readWorktreeMetaForHost(store, id, expectedHostId) ?? listed.ownedMeta(id) const worktree = { ...mergeWorktree(repo.id, gitWorktree, meta, repo.displayName), hostId: repoOwnerCount === 1 ? (meta?.hostId ?? expectedHostId) : expectedHostId @@ -210,6 +209,7 @@ export class RuntimeManagedWorktreeQueries { repoId: repo.id, authoritative: scan.ok && !scan.superseded, source: scan.ok ? 'git' : 'metadata-fallback', + workspaceMetadata: listed.workspaceMetadata, // Stamp the scan's starting generation: that is the catalog its rows describe. catalogVersion: localWorktreeCatalogVersionAt(scan.scanGeneration), worktrees: projectWorktreeLineageWithHostPeers(store, repo, detected) diff --git a/src/main/runtime/workspaces/catalog/runtime-workspace-change-events.ts b/src/main/runtime/workspaces/catalog/runtime-workspace-change-events.ts index c589d05f625..10a338ffee2 100644 --- a/src/main/runtime/workspaces/catalog/runtime-workspace-change-events.ts +++ b/src/main/runtime/workspaces/catalog/runtime-workspace-change-events.ts @@ -3,6 +3,7 @@ import type { AutomationsChangedPayload } from '../../../../shared/runtime-clien import { wakeFolderRepoGitUpgradeWatch } from '../../../ipc/folder-repo-git-upgrade-wake' import { runWorktreeChangeInvalidators } from '../../../ipc/worktrees/listing/worktree-change-invalidators' import type { WorktreeSortOrderUpdate } from '../../../../shared/worktree/sort-order-update' +import type { WorkspaceMetadataSet } from '../../../../shared/worktree/workspace-metadata-set' import { NativeChatLaunchDraftOwner } from '../../agent-session/launch/native-chat-launch-draft-owner' export abstract class RuntimeWorkspaceChangeEvents extends NativeChatLaunchDraftOwner { @@ -35,6 +36,12 @@ export abstract class RuntimeWorkspaceChangeEvents extends NativeChatLaunchDraft }) } + /** One workspace's metadata, addressed: clients patch that row instead of refetching its repo. Interim transport for the record. */ + publishWorkspaceMetadataSet(set: WorkspaceMetadataSet): void { + this.notifier?.workspaceMetadataSet?.(set) + this.emitClientEvent({ type: 'workspaceMetadataSet', ...set }) + } + /** Unlike coarse repoId-only changes, lifecycle events name the workspace; removals resolve only id/path before teardown. */ onWorktreeLifecycle(listener: (event: RuntimeWorktreeLifecycleEvent) => void): () => void { return this.worktreeLifecycleEvents.on(listener) diff --git a/src/main/startup/main-process-runtime-service.ts b/src/main/startup/main-process-runtime-service.ts index 7b0405af786..0306a69d3b1 100644 --- a/src/main/startup/main-process-runtime-service.ts +++ b/src/main/startup/main-process-runtime-service.ts @@ -1,4 +1,5 @@ import { wireRecipeWorkspaceRemoval } from '../ephemeral-vm/recipe-workspace-removal-wiring' +import { wireGitHubIssueLinkTitleFill } from '../github/github-issue-link-title-wiring' import { applySessionSearchSettingsChange, installChildSessionSearchService @@ -159,6 +160,9 @@ export function initializeMainProcessRuntime(): OrcaRuntimeService { recordDurableCrashBreadcrumb('agent_state_rules_active', rules) ) wireRecipeWorkspaceRemoval(runtime, store, () => app.getPath('userData')) + // Each host titles its own workspaces' untitled GitHub issue links once, when they are linked. + const disposeIssueLinkTitleFill = wireGitHubIssueLinkTitleFill(store, runtime) + app.once('will-quit', disposeIssueLinkTitleFill) state.runtime = runtime // Opening a recipe VM's workspace from a client wakes the VM through the runtime's one owner. registerRecipeVmWaker({ diff --git a/src/main/window/runtime-window-lifecycle.ts b/src/main/window/runtime-window-lifecycle.ts index b5d40e737b9..96b3d4f769d 100644 --- a/src/main/window/runtime-window-lifecycle.ts +++ b/src/main/window/runtime-window-lifecycle.ts @@ -45,6 +45,8 @@ export function registerRuntimeWindowLifecycle( } send('worktrees:changed', renamed ? { repoId, renamed } : { repoId }) }, + // Why: interim transport for the workspaceMetadata set record; the record shape, not this channel, is the contract. + workspaceMetadataSet: (set) => send('worktrees:metadataSet', set), worktreeBaseStatus: (event) => send('worktree:baseStatus', event), worktreeRemoteBranchConflict: (event) => send('worktree:remoteBranchConflict', event), reposChanged: () => send('repos:changed'), diff --git a/src/preload/api/worktree-api.ts b/src/preload/api/worktree-api.ts index 845020f4a67..c2cb7654a05 100644 --- a/src/preload/api/worktree-api.ts +++ b/src/preload/api/worktree-api.ts @@ -1,3 +1,4 @@ +import type { WorkspaceMetadataSet } from '../../shared/worktree/workspace-metadata-set' import type { SparsePresetFailure } from '../../shared/worktree/sparse-preset-failure' import type { ForgetRemovedWorktreesForExecutionHostArgs, @@ -157,6 +158,7 @@ export type WorktreeApi = { onHeadIdentitiesChanged: ( callback: (data: { repoId: string; identities: WorktreeHeadIdentity[] }) => void ) => () => void + onMetadataSet?: (callback: (set: WorkspaceMetadataSet) => void) => () => void onBaseStatus: (callback: (data: WorktreeBaseStatusEvent) => void) => () => void onRemoteBranchConflict: ( callback: (data: WorktreeRemoteBranchConflictEvent) => void diff --git a/src/preload/api/worktrees-bridge.ts b/src/preload/api/worktrees-bridge.ts index 7106dc7c4a3..9bbfb55fed1 100644 --- a/src/preload/api/worktrees-bridge.ts +++ b/src/preload/api/worktrees-bridge.ts @@ -9,6 +9,7 @@ import type { WorktreeHeadIdentity } from '../../shared/worktree/types' import type { PreloadApi } from '../api-types' +import type { WorkspaceMetadataSet } from '../../shared/worktree/workspace-metadata-set' export const worktreesApi = { listRetiredNames: (args) => ipcRenderer.invoke('worktrees:listRetiredNames', args), @@ -105,6 +106,12 @@ export const worktreesApi = { return () => ipcRenderer.removeListener('worktrees:headIdentitiesChanged', listener) }, + onMetadataSet: (callback: (set: WorkspaceMetadataSet) => void): (() => void) => { + const listener = (_event: Electron.IpcRendererEvent, set: WorkspaceMetadataSet) => callback(set) + ipcRenderer.on('worktrees:metadataSet', listener) + return () => ipcRenderer.removeListener('worktrees:metadataSet', listener) + }, + onBaseStatus: (callback: (data: WorktreeBaseStatusEvent) => void): (() => void) => { const listener = (_event: Electron.IpcRendererEvent, data: WorktreeBaseStatusEvent) => callback(data) diff --git a/src/renderer/src/hooks/ipc-events/project-catalog-ipc-bridge.ts b/src/renderer/src/hooks/ipc-events/project-catalog-ipc-bridge.ts index bb8c5e32b3f..8981c5646f4 100644 --- a/src/renderer/src/hooks/ipc-events/project-catalog-ipc-bridge.ts +++ b/src/renderer/src/hooks/ipc-events/project-catalog-ipc-bridge.ts @@ -1,3 +1,4 @@ +import { applyLocalWorktreeMetadataSet } from './worktree-metadata-set-events' import { applyWorktreeHeadIdentities } from '../worktree-head-identity-apply' import type { WorktreeChangeRefreshQueue } from '../worktree-change-refresh-queue' import { useAppStore } from '../../store' @@ -46,6 +47,10 @@ export function registerProjectCatalogIpcBridge( ) ) + if (window.api.worktrees.onMetadataSet) { + unsubs.push(window.api.worktrees.onMetadataSet(applyLocalWorktreeMetadataSet)) + } + if (window.api.worktrees.onHeadIdentitiesChanged) { unsubs.push( window.api.worktrees.onHeadIdentitiesChanged((data) => { diff --git a/src/renderer/src/hooks/ipc-events/runtime-client-ipc-bridge.ts b/src/renderer/src/hooks/ipc-events/runtime-client-ipc-bridge.ts index f3f855af18a..ea7bdb181c0 100644 --- a/src/renderer/src/hooks/ipc-events/runtime-client-ipc-bridge.ts +++ b/src/renderer/src/hooks/ipc-events/runtime-client-ipc-bridge.ts @@ -1,8 +1,8 @@ import '@/store/slices/linear/linear-issue-read' import { consumeLinearIssueRead } from '@/store/slices/linear/linear-issue-read-lifecycle' -import { applyHostWorktreeTerminalSleepState } from '@/components/terminal-pane/pty-transport/shutdown/pty-shutdown-exit-deferral' import { dispatchTerminalSideEffectBatch } from '@/components/terminal-pane/pty-transport/output/terminal-side-effect-facts-handler' import { emitAutomationsChangedWindowEvent } from '@/components/automations/events/automations-changed-window-event' +import { applyHostWorkspaceRowEvent } from './worktree-metadata-set-events' import { applyNativeChatLaunchDraftResolved } from '@/runtime/native-chat/launch-drafts/native-chat-launch-draft-runtime-resolution' import { getRuntimeEnvironmentRevision } from '@/runtime/rpc/environment-revisions/runtime-environment-revision' import { @@ -111,8 +111,7 @@ export function registerRuntimeClientIpcBridge( ) return } - if (event.type === 'worktreeTerminalSleepState') { - applyHostWorktreeTerminalSleepState(environmentId, event) + if (applyHostWorkspaceRowEvent(environmentId, event)) { return } if (event.type === 'terminalSideEffects') { diff --git a/src/renderer/src/hooks/ipc-events/worktree-metadata-set-events.ts b/src/renderer/src/hooks/ipc-events/worktree-metadata-set-events.ts new file mode 100644 index 00000000000..9ce76a4dc55 --- /dev/null +++ b/src/renderer/src/hooks/ipc-events/worktree-metadata-set-events.ts @@ -0,0 +1,98 @@ +import { applyHostWorktreeTerminalSleepState } from '@/components/terminal-pane/pty-transport/shutdown/pty-shutdown-exit-deferral' +import { + getRepoExecutionHostId, + toRuntimeExecutionHostId, + type ExecutionHostId +} from '../../../../shared/execution-host' +import type { RuntimeClientEvent } from '../../../../shared/runtime-client-events' +import { readWorkspaceMetadataSet } from '../../../../shared/worktree/workspace-metadata-set' +import { useAppStore } from '../../store' +import { + noteWorkspaceMetadataRecord, + requireWorkspaceMetadataCoverage, + workspaceMetadataProducerKey +} from '../../store/slices/worktrees/listing/workspace-metadata-coverage' +import { requestWorkspaceMetadataRecovery } from '../../store/slices/worktrees/listing/workspace-metadata-recovery' +import { + patchWorktreeMetadataSet, + type MetadataSetSource +} from '../../store/slices/worktrees/listing/worktree-metadata-set-patch' +import type { WorktreeFetchOptions } from '../../store/slices/worktrees/listing/worktree-fetch-options' + +const fetchWorktrees = (repoId: string, options: WorktreeFetchOptions): Promise => + useAppStore.getState().fetchWorktrees(repoId, options) + +function apply(record: unknown, source: MetadataSetSource): void { + const set = readWorkspaceMetadataSet(record) + if (!set) { + return + } + // The listing owner for this record's repo: this app (local and SSH rows) or the paired environment. + const listingHost: ExecutionHostId = + source.kind === 'environment' ? toRuntimeExecutionHostId(source.environmentId) : set.hostId + const producerKey = workspaceMetadataProducerKey(listingHost) + // Every repo projection this producer lists here, built only for a gap or a first-seen projection. + let projections: { repoId: string; hostId: ExecutionHostId }[] | undefined + let listed: Set | undefined + const producerProjections = (): { repoId: string; hostId: ExecutionHostId }[] => + (projections ??= useAppStore.getState().repos.flatMap((repo) => { + const hostId = getRepoExecutionHostId(repo) + return workspaceMetadataProducerKey(hostId) === producerKey + ? [{ repoId: repo.id, hostId }] + : [] + })) + const { replayed, gap } = noteWorkspaceMetadataRecord( + producerKey, + set.epoch, + set.seq, + set.repoId, + listingHost, + (repoId, hostId) => + (listed ??= new Set(producerProjections().map((p) => `${p.hostId}\u0000${p.repoId}`))).has( + `${hostId}\u0000${repoId}` + ) + ) + if (replayed) { + return + } + let applied = false + useAppStore.setState((current) => { + const result = patchWorktreeMetadataSet(current, set, source) + applied = result.applied + return result.patch ?? current + }) + if (gap) { + // A missed record names no projection: list every one this producer owns here, each once. + requireWorkspaceMetadataCoverage(producerKey, producerProjections()) + for (const { repoId, hostId } of producerProjections()) { + requestWorkspaceMetadataRecovery(repoId, hostId, fetchWorktrees) + } + } else if (!applied) { + // The row or a link isn't listed yet; a listing begun now includes this record. + requestWorkspaceMetadataRecovery(set.repoId, listingHost, fetchWorktrees) + } +} + +/** This app's own host: the record names the host its row lives on. */ +export function applyLocalWorktreeMetadataSet(record: unknown): void { + apply(record, { kind: 'local' }) +} + +/** + * A paired host's per-workspace row events, after the bridge's connection-generation check. + * Returns false for any other event. + */ +export function applyHostWorkspaceRowEvent( + environmentId: string, + event: RuntimeClientEvent +): boolean { + if (event.type === 'worktreeTerminalSleepState') { + applyHostWorktreeTerminalSleepState(environmentId, event) + return true + } + if (event.type === 'workspaceMetadataSet') { + apply(event, { kind: 'environment', environmentId }) + return true + } + return false +} diff --git a/src/renderer/src/lib/workspace/search/worktrees/worktree-palette-document.ts b/src/renderer/src/lib/workspace/search/worktrees/worktree-palette-document.ts index 5d62ca41274..62ee63ea12b 100644 --- a/src/renderer/src/lib/workspace/search/worktrees/worktree-palette-document.ts +++ b/src/renderer/src/lib/workspace/search/worktrees/worktree-palette-document.ts @@ -116,6 +116,18 @@ function resolveIssueTitle( return sources.issueCache[key]?.data?.title ?? '' } +// Each saved link is its own unit: same-number links (another repo, or the legacy issue field) must not +// shadow each other's saved titles. +function withEvidenceId(evidence: PaletteComposedEvidence, id: string): PaletteComposedEvidence { + const prefix = evidence.unit.id.length + return { + unit: { ...evidence.unit, id }, + fields: evidence.fields.map( + (field) => field && { ...field, id: `${id}${field.id.slice(prefix)}`, evidenceId: id } + ) + } +} + function buildEvidence( worktree: Worktree, repo: Repo | undefined, @@ -138,7 +150,7 @@ function buildEvidence( }), ...getWorkspaceAttachments({ linkedItems: worktree.linkedItems }).map((item) => { const evidence = buildWorktreeAttachmentEvidence(item) - return evidence ? { ...evidence, id: `attachment:${getWorkspaceAttachmentKey(item)}` } : null + return evidence && withEvidenceId(evidence, `attachment:${getWorkspaceAttachmentKey(item)}`) }), ...ports.map((port) => buildWorktreePortEvidence(port)) ] diff --git a/src/renderer/src/lib/workspace/search/worktrees/worktree-palette-search.test.ts b/src/renderer/src/lib/workspace/search/worktrees/worktree-palette-search.test.ts index 20dce1a770d..7ad40ac0666 100644 --- a/src/renderer/src/lib/workspace/search/worktrees/worktree-palette-search.test.ts +++ b/src/renderer/src/lib/workspace/search/worktrees/worktree-palette-search.test.ts @@ -125,4 +125,34 @@ describe('worktree-palette-search', () => { }) ]) }) + + it("finds each linked issue by its saved title before the issue's card has rendered", () => { + const issue = (number: number, title: string, url?: string) => ({ + provider: 'github' as const, + type: 'issue' as const, + number, + title, + ...(url ? { url } : {}) + }) + const worktrees = [ + makeWorktree({ + id: 'wt-cli', + linkedIssue: 11704, + linkedItems: [issue(11704, 'verificacion de permisos')] + }), + makeWorktree({ + id: 'wt-two-repos', + linkedIssue: 42, + linkedItems: [ + issue(42, 'alpha release', 'https://github.com/stablyai/orca/issues/42'), + issue(42, 'zebra crossing', 'https://github.com/other/repo/issues/42') + ] + }) + ] + const search = (query: string) => + searchWorktrees(worktrees, query, repoMap).map((result) => result.worktreeId) + + expect(search('verificacion')).toEqual(['wt-cli']) + expect(search('zebra')).toEqual(['wt-two-repos']) + }) }) diff --git a/src/renderer/src/runtime/rpc/subscriptions/runtime-client-events.ts b/src/renderer/src/runtime/rpc/subscriptions/runtime-client-events.ts index 48eecbbd24f..0c407253fc8 100644 --- a/src/renderer/src/runtime/rpc/subscriptions/runtime-client-events.ts +++ b/src/renderer/src/runtime/rpc/subscriptions/runtime-client-events.ts @@ -103,6 +103,7 @@ function isRuntimeClientEvent( isNotificationAttentionAcknowledgement(message.acknowledgement)) || message.type === 'reposChanged' || message.type === 'worktreesChanged' || + message.type === 'workspaceMetadataSet' || message.type === 'nativeChatLaunchDraftResolved' || message.type === 'terminalSideEffects' || message.type === 'sshStateChanged' || diff --git a/src/renderer/src/store/slices/worktrees/listing/detected-worktree-refresh-leases.ts b/src/renderer/src/store/slices/worktrees/listing/detected-worktree-refresh-leases.ts index e24c501bd51..ea21280af48 100644 --- a/src/renderer/src/store/slices/worktrees/listing/detected-worktree-refresh-leases.ts +++ b/src/renderer/src/store/slices/worktrees/listing/detected-worktree-refresh-leases.ts @@ -56,7 +56,8 @@ type ProviderInvocation = { export type DetectedWorktreeRefreshLeaseRegistry = { acquire( publicKey: string, - input: DetectedWorktreeRefreshProviderInput + input: DetectedWorktreeRefreshProviderInput, + fresh?: boolean ): DetectedWorktreeRefreshLease getActiveCounts(): { providerInvocations: number; waiterLeases: number } } @@ -198,15 +199,19 @@ export function createDetectedWorktreeRefreshLeaseRegistry( const acquire = ( publicKey: string, - input: DetectedWorktreeRefreshProviderInput + input: DetectedWorktreeRefreshProviderInput, + // A fresh listing must not join one that may have read metadata before a newer write. + fresh = false ): DetectedWorktreeRefreshLease => { - const invocation = - [...(invocationsByPublicKey.get(publicKey) ?? [])].find( - (candidate) => - !candidate.settled && - !candidate.cancelAttempted && - requestsAreCompatible(candidate.request, input) - ) ?? startInvocation(publicKey, input) + const joinable = fresh + ? undefined + : [...(invocationsByPublicKey.get(publicKey) ?? [])].find( + (candidate) => + !candidate.settled && + !candidate.cancelAttempted && + requestsAreCompatible(candidate.request, input) + ) + const invocation = joinable ?? startInvocation(publicKey, input) const waiterLeaseId = mintWaiterLeaseId() let resolveResult!: (result: HostQualifiedDetectedWorktreeResult) => void let rejectResult!: (error: unknown) => void diff --git a/src/renderer/src/store/slices/worktrees/listing/detected-worktree-refresh.ts b/src/renderer/src/store/slices/worktrees/listing/detected-worktree-refresh.ts index dc85da6755b..2987e36a5bd 100644 --- a/src/renderer/src/store/slices/worktrees/listing/detected-worktree-refresh.ts +++ b/src/renderer/src/store/slices/worktrees/listing/detected-worktree-refresh.ts @@ -61,10 +61,11 @@ export function acquireDetectedWorktreeRefreshLeaseForRepo( } const publicKey = detectedWorktreeRefreshKey(repoId, options) if (parsedHost.kind === 'local') { - return detectedWorktreeRefreshLeaseRegistry.acquire(publicKey, { - repoId, - executionHostId: LOCAL_EXECUTION_HOST_ID - }) + return detectedWorktreeRefreshLeaseRegistry.acquire( + publicKey, + { repoId, executionHostId: LOCAL_EXECUTION_HOST_ID }, + options.forceFresh + ) } if ( !options.directSshAuthority || @@ -72,11 +73,15 @@ export function acquireDetectedWorktreeRefreshLeaseForRepo( ) { throw new Error('Direct SSH provider leases require exact target authority') } - return detectedWorktreeRefreshLeaseRegistry.acquire(publicKey, { - repoId, - executionHostId: options.executionHostId as SshExecutionHostId, - expectedAuthority: { ...options.directSshAuthority } - }) + return detectedWorktreeRefreshLeaseRegistry.acquire( + publicKey, + { + repoId, + executionHostId: options.executionHostId as SshExecutionHostId, + expectedAuthority: { ...options.directSshAuthority } + }, + options.forceFresh + ) } export function qualifiedProviderResultIsAdmitted( @@ -148,7 +153,7 @@ async function listDetectedWorktreesForRuntimeRepoOnce( const key = detectedWorktreeRefreshKey(repoId, options) const connectionGeneration = getEnvironmentSshStateGeneration(environmentId) const runtimeConnectionGeneration = getRuntimeEnvironmentConnectionGeneration(environmentId) - let refresh = runtimeDetectedWorktreeRefreshesInFlight.get(key) + let refresh = options.forceFresh ? undefined : runtimeDetectedWorktreeRefreshesInFlight.get(key) if (!refresh) { refresh = listDetectedWorktreesForRepo({ kind: 'environment', environmentId }, repoId, { reuseRecentCompatibilityFailure: options.reuseRecentCompatibilityFailure diff --git a/src/renderer/src/store/slices/worktrees/listing/fetch-all-worktrees.ts b/src/renderer/src/store/slices/worktrees/listing/fetch-all-worktrees.ts index ac72a2588c0..bbe2405acac 100644 --- a/src/renderer/src/store/slices/worktrees/listing/fetch-all-worktrees.ts +++ b/src/renderer/src/store/slices/worktrees/listing/fetch-all-worktrees.ts @@ -13,6 +13,7 @@ import { } from './worktree-host-ownership' import { fetchKnownSshWorktreesForRepo } from './known-ssh-worktree-fetch' import { mergeFetchedWorktrees } from './fetched-worktree-merge' +import { listingMetadataCoverage } from './workspace-metadata-coverage' import { notifyRuntimeScopeForbiddenIfNeeded } from './runtime-scope-forbidden-toast' import { mapReposForWorktreeRefresh } from './worktree-refresh-pool' import { selectWorktreesByRepo } from '../../../workspaces/catalog/workspace-record-fields' @@ -64,15 +65,19 @@ export function createFetchAllWorktrees( if (refresh.status !== 'admitted') { return failed } + const metadata = listingMetadataCoverage(get(), repo.id, hostId, refresh) const outcome = mergeFetchedWorktrees(set, get, { repoId: repo.id, hostId, ownerWasMissingAtStart: false, requestStartedWorktrees, setup, - refresh, + refresh: metadata.refresh, purgeRemovedWorktrees }) + if (outcome === 'applied') { + metadata.settle((id, fetchOptions) => get().fetchWorktrees(id, fetchOptions)) + } return outcome === 'applied' ? { repoId: repo.id, hostId, refresh } : failed } catch (err) { if (!notifyRuntimeScopeForbiddenIfNeeded(err)) { diff --git a/src/renderer/src/store/slices/worktrees/listing/fetch-worktrees.ts b/src/renderer/src/store/slices/worktrees/listing/fetch-worktrees.ts index 6c42355d76d..4761ad69ff3 100644 --- a/src/renderer/src/store/slices/worktrees/listing/fetch-worktrees.ts +++ b/src/renderer/src/store/slices/worktrees/listing/fetch-worktrees.ts @@ -20,6 +20,7 @@ import { } from './worktree-host-ownership' import { fetchKnownSshWorktreesForRepo } from './known-ssh-worktree-fetch' import { mergeFetchedWorktrees } from './fetched-worktree-merge' +import { listingMetadataCoverage } from './workspace-metadata-coverage' import { notifyRuntimeScopeForbiddenIfNeeded } from './runtime-scope-forbidden-toast' import { refreshRemoteWorktreeLineageBestEffort } from '../metadata/worktree-lineage-refresh' import { selectWorktreesByRepo } from '../../../workspaces/catalog/workspace-record-fields' @@ -87,7 +88,8 @@ export function createFetchWorktrees( connectionId: repoOwner?.connectionId, knownWorktreeIds: getKnownWorktreeIdsForPurge(ownerState, repoId, hostId), isStaleCatalogPublication: (result) => - isStaleWorktreeCatalogPublication(get(), repoId, hostId, result.catalogVersion) + isStaleWorktreeCatalogPublication(get(), repoId, hostId, result.catalogVersion), + ...(options?.forceFresh ? { forceFresh: true } : {}) }) if (refresh.status !== 'admitted') { return directCallerAuthority ? refresh.providerResult : false @@ -95,6 +97,7 @@ export function createFetchWorktrees( if (options?.requireAuthoritative && !refresh.result.authoritative) { return directCallerAuthority ? (refresh.providerResult ?? false) : false } + const metadata = listingMetadataCoverage(get(), repoId, hostId, refresh) const outcome = mergeFetchedWorktrees(set, get, { repoId, hostId, @@ -105,8 +108,11 @@ export function createFetchWorktrees( : undefined, requestStartedWorktrees, setup, - refresh + refresh: metadata.refresh }) + if (outcome === 'applied') { + metadata.settle((id, fetchOptions) => get().fetchWorktrees(id, fetchOptions)) + } switch (outcome) { case 'applied': break diff --git a/src/renderer/src/store/slices/worktrees/listing/workspace-metadata-coverage.test.ts b/src/renderer/src/store/slices/worktrees/listing/workspace-metadata-coverage.test.ts new file mode 100644 index 00000000000..6ae672a6ff9 --- /dev/null +++ b/src/renderer/src/store/slices/worktrees/listing/workspace-metadata-coverage.test.ts @@ -0,0 +1,76 @@ +import { describe, expect, it, vi } from 'vitest' +import { + noteWorkspaceMetadataRecord, + requireWorkspaceMetadataCoverage, + workspaceMetadataListingCoverage +} from './workspace-metadata-coverage' +import { requestWorkspaceMetadataRecovery } from './workspace-metadata-recovery' + +const listed = () => true + +describe('workspace metadata listing coverage', () => { + it('a missed record makes every repo of that producer wait for a listing that includes it', () => { + expect(noteWorkspaceMetadataRecord('local', 'gap', 1, 'repo-a', 'local', listed)).toEqual({ + replayed: false, + gap: false + }) + expect(noteWorkspaceMetadataRecord('local', 'gap', 3, 'repo-a', 'local', listed).gap).toBe(true) + requireWorkspaceMetadataCoverage('local', [ + { repoId: 'repo-a', hostId: 'local' }, + { repoId: 'repo-b', hostId: 'ssh:box' } + ]) + + expect( + workspaceMetadataListingCoverage('local', 'repo-b', 'ssh:box', { epoch: 'gap', seq: 2 }) + ).toBe('stale') + expect( + workspaceMetadataListingCoverage('local', 'repo-b', 'ssh:box', { epoch: 'gap', seq: 3 }) + ).toBe('covered') + // The same repo listed on another host is a different projection. + expect( + workspaceMetadataListingCoverage('local', 'repo-b', 'local', { epoch: 'gap', seq: 0 }) + ).toBe('unknown') + + // A later gap after repo-b was removed forgets it. + noteWorkspaceMetadataRecord('local', 'gap', 5, 'repo-a', 'local', listed) + requireWorkspaceMetadataCoverage('local', [{ repoId: 'repo-a', hostId: 'local' }]) + expect( + workspaceMetadataListingCoverage('local', 'repo-b', 'ssh:box', { epoch: 'gap', seq: 5 }) + ).toBe('unknown') + }) + + it("another producer run's listing keeps titles without claiming to cover them", () => { + noteWorkspaceMetadataRecord('env:e1:1:1', 'before', 9, 'repo-a', 'runtime:e1', listed) + expect( + workspaceMetadataListingCoverage('env:e1:1:1', 'repo-a', 'runtime:e1', { + epoch: 'after', + seq: 1 + }) + ).toBe('unstamped') + }) +}) + +describe('workspace metadata recovery listing', () => { + it('runs one fresh listing at a time and once more for requests made during it', async () => { + let finish: () => void = () => {} + const fetchWorktrees = vi.fn( + () => new Promise((resolve) => (finish = () => resolve(true))) + ) + + requestWorkspaceMetadataRecovery('repo-a', 'local', fetchWorktrees) + requestWorkspaceMetadataRecovery('repo-a', 'local', fetchWorktrees) + requestWorkspaceMetadataRecovery('repo-a', 'local', fetchWorktrees) + expect(fetchWorktrees).toHaveBeenCalledTimes(1) + expect(fetchWorktrees).toHaveBeenCalledWith('repo-a', { + forceLocalOwner: true, + forceFresh: true, + suppressRemoteLineageRefresh: true + }) + + finish() + await vi.waitFor(() => expect(fetchWorktrees).toHaveBeenCalledTimes(2)) + finish() + await new Promise((resolve) => setTimeout(resolve, 0)) + expect(fetchWorktrees).toHaveBeenCalledTimes(2) + }) +}) diff --git a/src/renderer/src/store/slices/worktrees/listing/workspace-metadata-coverage.ts b/src/renderer/src/store/slices/worktrees/listing/workspace-metadata-coverage.ts new file mode 100644 index 00000000000..f43e4b932b6 --- /dev/null +++ b/src/renderer/src/store/slices/worktrees/listing/workspace-metadata-coverage.ts @@ -0,0 +1,235 @@ +import { parseExecutionHostId, type ExecutionHostId } from '../../../../../../shared/execution-host' +import type { Worktree } from '../../../../../../shared/worktree/types' +import type { WorkspaceMetadataCursor } from '../../../../../../shared/worktree/workspace-metadata-set' +import { getWorkspaceAttachments } from '../../../../../../shared/workspace-attachments/workspace-attachments' +import { getWorkspaceReferenceIdentity } from '../../../../../../shared/workspace-attachments/workspace-reference-identity' +import { + getRuntimeEnvironmentRevision, + onRuntimeEnvironmentsRetired +} from '../../../../runtime/rpc/environment-revisions/runtime-environment-revision' +import { getRuntimeEnvironmentConnectionGeneration } from '../../runtime-status/runtime-status-connection-generation' +import type { AppState } from '../../../types' +import { selectWorktreesByRepo } from '../../../workspaces/catalog/workspace-record-fields' +import { worktreeHostMatchOptions, worktreeMatchesHost } from './worktree-host-ownership' +import type { AdmittedDetectedWorktreeRefresh } from './worktree-slice-types' +import type { WorktreeFetchOptions } from './worktree-fetch-options' +import { requestWorkspaceMetadataRecovery } from './workspace-metadata-recovery' + +// One producer's workspaceMetadata stream as this client has seen it: the last record seq, and per listed +// projection (repo on a host) the newest record that touched it. +type Projection = { repoId: string; hostId: ExecutionHostId; latest: number } +type ProducerCoverage = { epoch: string; lastSeq: number; projections: Map } + +const producers = new Map() + +// Encoded: environment ids (WSL ones) may contain ':', which would let one id prefix another's keys. +const environmentKeyPrefix = (environmentId: string): string => + `env:${encodeURIComponent(environmentId)}:` + +function dropEnvironmentProducers(environmentId: string, except?: string): void { + for (const key of producers.keys()) { + if (key !== except && key.startsWith(environmentKeyPrefix(environmentId))) { + producers.delete(key) + } + } +} + +// A removed or re-pointed environment's stream ends with it. +onRuntimeEnvironmentsRetired((environmentIds) => { + for (const environmentId of environmentIds) { + dropEnvironmentProducers(environmentId) + } +}) + +const projectionKey = (repoId: string, hostId: ExecutionHostId): string => + `${hostId}\u0000${repoId}` + +/** + * The producer whose cursor a host's listings and records share: this app's own main for local and + * SSH rows, or a paired environment for its rows, keyed by that pairing and connection. + */ +export function workspaceMetadataProducerKey(hostId: ExecutionHostId): string { + const parsed = parseExecutionHostId(hostId) + if (parsed?.kind !== 'runtime') { + return 'local' + } + const environmentId = parsed.environmentId + return `${environmentKeyPrefix(environmentId)}${getRuntimeEnvironmentRevision(environmentId) ?? ''}:${getRuntimeEnvironmentConnectionGeneration(environmentId)}` +} + +function producerFor(key: string, epoch: string, hostId: ExecutionHostId): ProducerCoverage { + const current = producers.get(key) + if (current?.epoch === epoch) { + return current + } + // A new epoch is a restarted producer: what it lists already includes its earlier writes. + const fresh = { epoch, lastSeq: 0, projections: new Map() } + const parsed = parseExecutionHostId(hostId) + if (parsed?.kind === 'runtime') { + // A reconnected environment is a new producer; its earlier bookkeeping dies here. + dropEnvironmentProducers(parsed.environmentId, key) + } + producers.set(key, fresh) + return fresh +} + +/** + * Notes a record before it is applied. `replayed` records are dropped; a `gap` means some earlier + * record was missed and its projection is unknown, so every projection of this producer needs a + * covering listing. A projection seen for the first time forgets those no longer listed here. + */ +export function noteWorkspaceMetadataRecord( + key: string, + epoch: string, + seq: number, + repoId: string, + hostId: ExecutionHostId, + isListed: (repoId: string, hostId: ExecutionHostId) => boolean +): { replayed: boolean; gap: boolean } { + const producer = producerFor(key, epoch, hostId) + if (seq <= producer.lastSeq) { + return { replayed: true, gap: false } + } + const gap = producer.lastSeq > 0 && seq > producer.lastSeq + 1 + producer.lastSeq = seq + const recorded = projectionKey(repoId, hostId) + if (!producer.projections.has(recorded)) { + for (const [projection, entry] of producer.projections) { + if (!isListed(entry.repoId, entry.hostId)) { + producer.projections.delete(projection) + } + } + } + producer.projections.set(recorded, { repoId, hostId, latest: seq }) + return { replayed: false, gap } +} + +/** + * After a gap, every listed projection of the producer must be listed again at or past the last seq; + * the live projections replace all earlier ones. + */ +export function requireWorkspaceMetadataCoverage( + key: string, + projections: Iterable<{ repoId: string; hostId: ExecutionHostId }> +): void { + const producer = producers.get(key) + if (!producer) { + return + } + producer.projections = new Map( + [...projections].map(({ repoId, hostId }) => [ + projectionKey(repoId, hostId), + { repoId, hostId, latest: producer.lastSeq } + ]) + ) +} + +/** + * Where a listing stands against the newest record that touched its projection. `stale` read before + * that record; `unstamped` comes from a path, host or producer run that can't say; `unknown` has nothing to check. + */ +export function workspaceMetadataListingCoverage( + key: string, + repoId: string, + hostId: ExecutionHostId, + stamp: WorkspaceMetadataCursor | undefined +): 'covered' | 'stale' | 'unstamped' | 'unknown' { + const producer = producers.get(key) + const projection = producer?.projections.get(projectionKey(repoId, hostId)) + if (!producer || !projection) { + return 'unknown' + } + // Another epoch's stamp can't be ordered against this one's records (a restart either way). + if (!stamp || stamp.epoch !== producer.epoch) { + return 'unstamped' + } + return stamp.seq >= projection.latest ? 'covered' : 'stale' +} + +/** + * A listing that read metadata before a published write must not erase what that write already showed: + * keep the current rows' issue titles on the same, still-untitled links. Never adds or restores a link. + */ +function keepPublishedLinkTitles( + incoming: T[], + current: readonly Worktree[] +): T[] { + const currentById = new Map(current.map((row) => [row.id, row])) + return incoming.map((row) => { + const previous = currentById.get(row.id) + if (!previous) { + return row + } + const titles = new Map() + for (const item of getWorkspaceAttachments(previous)) { + const title = item.title?.trim() + if (item.provider === 'github' && item.type === 'issue' && title) { + titles.set(getWorkspaceReferenceIdentity(item), title) + } + } + if (titles.size === 0) { + return row + } + let kept = false + const linkedItems = getWorkspaceAttachments(row).map((item) => { + const title = item.title?.trim() ? undefined : titles.get(getWorkspaceReferenceIdentity(item)) + if (!title) { + return item + } + kept = true + return { ...item, title } + }) + return kept ? { ...row, linkedItems } : row + }) +} + +function currentRows(state: AppState, repoId: string, hostId: ExecutionHostId): Worktree[] { + const matchOptions = worktreeHostMatchOptions(state, repoId, hostId) + return (selectWorktreesByRepo(state)[repoId] ?? []).filter((row) => + worktreeMatchesHost(row, hostId, matchOptions) + ) +} + +type FetchWorktrees = (repoId: string, options: WorktreeFetchOptions) => Promise + +/** + * A listing's place against the newest record for its repo and host, checked right before it is + * merged: a stale listing keeps already-shown titles and, once applied, asks for a fresh listing. + */ +export function listingMetadataCoverage( + state: AppState, + repoId: string, + hostId: ExecutionHostId, + refresh: AdmittedDetectedWorktreeRefresh +): { refresh: AdmittedDetectedWorktreeRefresh; settle: (fetchWorktrees: FetchWorktrees) => void } { + const key = workspaceMetadataProducerKey(hostId) + const coverage = workspaceMetadataListingCoverage( + key, + repoId, + hostId, + refresh.result.workspaceMetadata + ) + // Keeping titles already shown is always safe (it never adds or restores a link), so any listing that + // can't prove it includes the newest record keeps them. + const keep = coverage === 'stale' || coverage === 'unstamped' + return { + refresh: keep + ? { + ...refresh, + result: { + ...refresh.result, + worktrees: keepPublishedLinkTitles( + refresh.result.worktrees, + currentRows(state, repoId, hostId) + ) + } + } + : refresh, + settle: (fetchWorktrees) => { + // Whatever the stale listing dropped (a newer link, say), a listing begun now includes. + if (coverage === 'stale') { + requestWorkspaceMetadataRecovery(repoId, hostId, fetchWorktrees) + } + } + } +} diff --git a/src/renderer/src/store/slices/worktrees/listing/workspace-metadata-recovery.ts b/src/renderer/src/store/slices/worktrees/listing/workspace-metadata-recovery.ts new file mode 100644 index 00000000000..938afa9f88c --- /dev/null +++ b/src/renderer/src/store/slices/worktrees/listing/workspace-metadata-recovery.ts @@ -0,0 +1,40 @@ +import { + LOCAL_EXECUTION_HOST_ID, + type ExecutionHostId +} from '../../../../../../shared/execution-host' +import type { WorktreeFetchOptions } from './worktree-fetch-options' + +type FetchWorktrees = (repoId: string, options: WorktreeFetchOptions) => Promise + +// One recovery listing per repo and host at a time; a request during one runs once more after it. +const inFlight = new Set() +const rerun = new Set() + +/** + * Lists a repo afresh (never joining a listing that may predate a published metadata write). The + * listing's own coverage check asks again if it still turns out stale. + */ +export function requestWorkspaceMetadataRecovery( + repoId: string, + hostId: ExecutionHostId, + fetchWorktrees: FetchWorktrees +): void { + const key = `${hostId}\u0000${repoId}` + if (inFlight.has(key)) { + rerun.add(key) + return + } + inFlight.add(key) + const options: WorktreeFetchOptions = + hostId === LOCAL_EXECUTION_HOST_ID + ? { forceLocalOwner: true, forceFresh: true, suppressRemoteLineageRefresh: true } + : { executionHostId: hostId, forceFresh: true, suppressRemoteLineageRefresh: true } + void fetchWorktrees(repoId, options) + .catch(() => undefined) + .finally(() => { + inFlight.delete(key) + if (rerun.delete(key)) { + requestWorkspaceMetadataRecovery(repoId, hostId, fetchWorktrees) + } + }) +} diff --git a/src/renderer/src/store/slices/worktrees/listing/worktree-fetch-options.ts b/src/renderer/src/store/slices/worktrees/listing/worktree-fetch-options.ts index a4ba5fb93a8..f3375da59b0 100644 --- a/src/renderer/src/store/slices/worktrees/listing/worktree-fetch-options.ts +++ b/src/renderer/src/store/slices/worktrees/listing/worktree-fetch-options.ts @@ -9,6 +9,8 @@ export type WorktreeFetchOptions = { forceLocalOwner?: boolean /** Skip automatic remote lineage when the caller owns a final host-wide refresh. */ suppressRemoteLineageRefresh?: boolean + /** Start a new listing instead of joining one in flight (it may predate a published metadata write). */ + forceFresh?: boolean } export type DirectSshWorktreeFetchOptions = WorktreeFetchOptions & { diff --git a/src/renderer/src/store/slices/worktrees/listing/worktree-metadata-set-patch.test.ts b/src/renderer/src/store/slices/worktrees/listing/worktree-metadata-set-patch.test.ts new file mode 100644 index 00000000000..4af41008656 --- /dev/null +++ b/src/renderer/src/store/slices/worktrees/listing/worktree-metadata-set-patch.test.ts @@ -0,0 +1,124 @@ +import { describe, expect, it, vi } from 'vitest' +import type { WorkspaceAttachment } from '../../../../../../shared/worktree/types' +import type { AppState } from '../../../types' +import { useAppStore } from '../../..' +import { + makeRepo, + makeWorktree +} from '../../../../components/worktree-jump/worktree-jump-palette-test-fixtures' +import { selectWorktreesByRepo } from '../../../workspaces/catalog/workspace-record-fields' +import { patchWorktreeMetadataSet } from './worktree-metadata-set-patch' +import { listingMetadataCoverage, noteWorkspaceMetadataRecord } from './workspace-metadata-coverage' +import type { AdmittedDetectedWorktreeRefresh } from './worktree-slice-types' + +const ID = 'repo-1::/work/orca' +const issue = (title?: string): WorkspaceAttachment[] => [ + { provider: 'github', type: 'issue', number: 42, ...(title ? { title } : {}) } +] +const local = { kind: 'local' as const } + +function state(linkedItems: WorkspaceAttachment[] = issue()): AppState { + const linkedIssue = linkedItems.length > 0 ? 42 : null + return { + ...useAppStore.getInitialState(), + repos: [makeRepo()], + worktreesByRepo: { + 'repo-1': [ + makeWorktree(ID, 'local orca', { hostId: 'local', linkedIssue, linkedItems }), + makeWorktree(ID, 'ssh orca', { hostId: 'ssh:build-box', linkedIssue, linkedItems }) + ] + } + } +} + +function record(seq: number, epoch: string, hostId: 'local' | 'ssh:build-box' = 'ssh:build-box') { + return { + kind: 'set' as const, + domain: 'workspaceMetadata' as const, + hostId, + epoch, + seq, + id: `${hostId}::${ID}`, + repoId: 'repo-1', + worktreeId: ID, + row: { + linkedItems: issue('Fix login'), + linkedWorkItem: null, + linkedIssue: 42, + linkedTaskSourceContext: null + } + } +} + +const linksOn = (value: AppState, hostId: string) => + selectWorktreesByRepo(value)['repo-1'].find((row) => row.hostId === hostId)?.linkedItems + +describe('addressed workspace metadata records', () => { + it('title only the named host row, and never bring back an unlinked issue', () => { + const { patch, applied } = patchWorktreeMetadataSet(state(), record(1, 'e1'), local) + const titled = { ...state(), ...patch } + expect(applied).toBe(true) + expect(linksOn(titled, 'ssh:build-box')).toEqual(issue('Fix login')) + expect(linksOn(titled, 'local')).toEqual(issue()) + + expect(patchWorktreeMetadataSet(state([]), record(2, 'e1'), local)).toEqual({ + patch: null, + applied: false + }) + + // A second titled link the row doesn't list yet still needs a listing. + const twoLinks = record(3, 'e1') + twoLinks.row.linkedItems = [ + ...issue('Fix login'), + { provider: 'github', type: 'issue', number: 43, title: 'Fix logout' } + ] + expect(patchWorktreeMetadataSet(state(), twoLinks, local).applied).toBe(false) + }) + + it('keeps a shown title through a listing that predates it, then lists the repo afresh', async () => { + noteWorkspaceMetadataRecord('local', 'e2', 7, 'repo-1', 'local', () => true) + const shown = { + ...state(), + ...patchWorktreeMetadataSet(state(), record(7, 'e2', 'local'), local).patch + } + const listing = (seq: number): AdmittedDetectedWorktreeRefresh => ({ + status: 'admitted', + executionHostId: 'local', + result: { + repoId: 'repo-1', + authoritative: true, + source: 'git', + workspaceMetadata: { epoch: 'e2', seq }, + worktrees: [ + { + ...makeWorktree(ID, 'local orca', { linkedIssue: 42, linkedItems: issue() }), + ownership: 'orca-managed', + selectedCheckout: false, + visible: true + } + ] + } + }) + const fetchWorktrees = vi.fn(async () => true) + + const stale = listingMetadataCoverage(shown, 'repo-1', 'local', listing(6)) + expect(stale.refresh.result.worktrees[0].linkedItems).toEqual(issue('Fix login')) + stale.settle(fetchWorktrees) + expect(fetchWorktrees).toHaveBeenCalledWith('repo-1', { + forceLocalOwner: true, + forceFresh: true, + suppressRemoteLineageRefresh: true + }) + + const covering = listingMetadataCoverage(shown, 'repo-1', 'local', listing(7)) + expect(covering.refresh.result.worktrees[0].linkedItems).toEqual(issue()) + covering.settle(fetchWorktrees) + expect(fetchWorktrees).toHaveBeenCalledTimes(1) + // A listing that read before record 7 and lands after the covering one still keeps the title, and + // lists again for anything else it dropped. + const late = listingMetadataCoverage(shown, 'repo-1', 'local', listing(6)) + expect(late.refresh.result.worktrees[0].linkedItems).toEqual(issue('Fix login')) + late.settle(fetchWorktrees) + await vi.waitFor(() => expect(fetchWorktrees).toHaveBeenCalledTimes(2)) + }) +}) diff --git a/src/renderer/src/store/slices/worktrees/listing/worktree-metadata-set-patch.ts b/src/renderer/src/store/slices/worktrees/listing/worktree-metadata-set-patch.ts new file mode 100644 index 00000000000..fd89ff6bc07 --- /dev/null +++ b/src/renderer/src/store/slices/worktrees/listing/worktree-metadata-set-patch.ts @@ -0,0 +1,83 @@ +import { + parseExecutionHostId, + toRuntimeExecutionHostId +} from '../../../../../../shared/execution-host' +import type { Worktree } from '../../../../../../shared/worktree/types' +import type { WorkspaceMetadataSet } from '../../../../../../shared/worktree/workspace-metadata-set' +import { getWorkspaceAttachments } from '../../../../../../shared/workspace-attachments/workspace-attachments' +import { getWorkspaceReferenceIdentity } from '../../../../../../shared/workspace-attachments/workspace-reference-identity' +import type { AppState } from '../../../types' +import { worktreeHostMatchOptions, worktreeMatchesHost } from './worktree-host-ownership' +import { selectWorktreesByRepo } from '../../../workspaces/catalog/workspace-record-fields' + +/** Where a record came from: this app's own main, or a paired host's runtime event stream. */ +export type MetadataSetSource = { kind: 'local' } | { kind: 'environment'; environmentId: string } + +// The row a record names, by its own execution host; a paired host's SSH twin is a different row. +function rowMatcher( + state: AppState, + set: WorkspaceMetadataSet, + source: MetadataSetSource +): (row: Worktree) => boolean { + if (source.kind === 'local') { + const options = worktreeHostMatchOptions(state, set.repoId, set.hostId) + return (row) => row.id === set.worktreeId && worktreeMatchesHost(row, set.hostId, options) + } + const environmentHost = toRuntimeExecutionHostId(source.environmentId) + const options = worktreeHostMatchOptions(state, set.repoId, environmentHost) + const sshHost = parseExecutionHostId(set.hostId)?.kind === 'ssh' ? set.hostId : null + return (row) => + row.id === set.worktreeId && + worktreeMatchesHost(row, environmentHost, options) && + (sshHost ? row.hostId === sshHost : parseExecutionHostId(row.hostId ?? null)?.kind !== 'ssh') +} + +/** + * Fills a producer's addressed `workspaceMetadata` record into the row it names. The only producer + * today saves issue titles, so only still-untitled links with the same identity take a title; the + * record's links never replace the row's. `applied` is false when the row or link isn't listed yet. + */ +export function patchWorktreeMetadataSet( + state: AppState, + set: WorkspaceMetadataSet, + source: MetadataSetSource +): { patch: Pick | null; applied: boolean } { + const titles = new Map() + for (const item of set.row.linkedItems ?? []) { + const title = item.title?.trim() + if (item.provider === 'github' && item.type === 'issue' && title) { + titles.set(getWorkspaceReferenceIdentity(item), title) + } + } + const rows = selectWorktreesByRepo(state)[set.repoId] + const index = titles.size > 0 && rows ? rows.findIndex(rowMatcher(state, set, source)) : -1 + if (!rows || index === -1) { + return { patch: null, applied: titles.size === 0 } + } + const seen = new Set() + let filled = false + const linkedItems = getWorkspaceAttachments(rows[index]).map((item) => { + const identity = getWorkspaceReferenceIdentity(item) + const title = titles.get(identity) + if (!title) { + return item + } + seen.add(identity) + if (item.title?.trim()) { + return item + } + filled = true + return { ...item, title } + }) + // Every titled link must be listed here; one that isn't needs a listing that includes it. + const applied = seen.size === titles.size + if (!filled) { + return { patch: null, applied } + } + const next = rows.slice() + next[index] = { ...rows[index], linkedItems } + return { + patch: { worktreesByRepo: { ...selectWorktreesByRepo(state), [set.repoId]: next } }, + applied + } +} diff --git a/src/renderer/src/store/slices/worktrees/listing/worktree-slice-types.ts b/src/renderer/src/store/slices/worktrees/listing/worktree-slice-types.ts index e4094088a09..8abc30b2011 100644 --- a/src/renderer/src/store/slices/worktrees/listing/worktree-slice-types.ts +++ b/src/renderer/src/store/slices/worktrees/listing/worktree-slice-types.ts @@ -34,6 +34,8 @@ export type DetectedWorktreeRefreshOptions = BackgroundRuntimeRefreshOptions & { // Why a callback: staleness is judged against the client's live applied version, which a // create or remove reply can advance while the listing is in flight. isStaleCatalogPublication?: (result: DetectedWorktreeListResult) => boolean + // Start a new listing instead of joining one in flight (it may predate a published metadata write). + forceFresh?: boolean } export type AdmittedDetectedWorktreeRefresh = { diff --git a/src/renderer/src/store/slices/worktrees/metadata/update-worktree-meta.ts b/src/renderer/src/store/slices/worktrees/metadata/update-worktree-meta.ts index 77d3c6969e9..b015759b24e 100644 --- a/src/renderer/src/store/slices/worktrees/metadata/update-worktree-meta.ts +++ b/src/renderer/src/store/slices/worktrees/metadata/update-worktree-meta.ts @@ -120,7 +120,8 @@ export function createUpdateWorktreeMeta( return { ok: true } } normalizedUpdates = normalizeGitHubPRSuppressionUpdate( - normalizeWorkspaceAttachmentUpdate(worktreeForUpdate, mutationUpdates) + normalizeWorkspaceAttachmentUpdate(worktreeForUpdate, mutationUpdates), + worktreeForUpdate?.linkedPR ) const currentResolvedPushTarget = normalizedUpdates.linkedPR === linkedPrForPushTarget ? resolvedPushTarget : undefined diff --git a/src/renderer/src/web/preload-api/web-worktrees-api.ts b/src/renderer/src/web/preload-api/web-worktrees-api.ts index 85a06a72a75..961e52ceb23 100644 --- a/src/renderer/src/web/preload-api/web-worktrees-api.ts +++ b/src/renderer/src/web/preload-api/web-worktrees-api.ts @@ -267,6 +267,7 @@ export function createWorktreesApi(): NonNullable['worktrees onChanged: () => noopUnsubscribe, onGitStatusMetadataChanged: () => noopUnsubscribe, onHeadIdentitiesChanged: () => noopUnsubscribe, + onMetadataSet: () => noopUnsubscribe, onBaseStatus: () => noopUnsubscribe, onRemoteBranchConflict: () => noopUnsubscribe } diff --git a/src/shared/runtime-client-events.ts b/src/shared/runtime-client-events.ts index 65cda9b9c1c..f12d57daf00 100644 --- a/src/shared/runtime-client-events.ts +++ b/src/shared/runtime-client-events.ts @@ -10,6 +10,7 @@ import type { TerminalSideEffectBatch } from './terminal/output/terminal-side-ef import type { RuntimeNativeChatLaunchDraftResolution } from './runtime-types' import type { RuntimeNavigationTarget } from './runtime-navigation' import type { WorktreeSortOrderUpdate } from './worktree/sort-order-update' +import type { WorkspaceMetadataSet } from './worktree/workspace-metadata-set' import type { SidebarInputsChangedEvent } from './sidebar-list/sidebar-inputs-feed-contract' import type { BrowserClientHostRequestEvent } from './browser/client-hosting/browser-client-host-request-protocol' @@ -38,6 +39,8 @@ export type RuntimeClientEvent = } /** An agent row changed (coalesced, no ids); phones re-read worktree.ps. Sent only while the sidebar feed is off. */ | { type: 'worktreeRowsChanged' } + // Why: addressed workspace metadata; old clients drop the unknown type and see it on their next listing. + | ({ type: 'workspaceMetadataSet' } & WorkspaceMetadataSet) /** What mobileRelay.hosts.list answers changed; phones re-read it instead of polling. */ | { type: 'relayHostsChanged' } /** A desktop PR refresh found this checkout's PR (or its absence) changed; phones re-read it. */ diff --git a/src/shared/worktree/github-pr-suppression.ts b/src/shared/worktree/github-pr-suppression.ts index f89ccbfc8e2..e7d671e59ce 100644 --- a/src/shared/worktree/github-pr-suppression.ts +++ b/src/shared/worktree/github-pr-suppression.ts @@ -10,10 +10,14 @@ export function isGitHubPRSuppressed( return linkedPR === null && suppressedGitHubPR === prNumber } +/** Selecting a PR lifts a suppression; restating the already-selected PR leaves it alone. */ export function normalizeGitHubPRSuppressionUpdate( - updates: Partial + updates: Partial, + currentLinkedPR?: number | null ): Partial { - return typeof updates.linkedPR === 'number' && updates.linkedPR > 0 + return typeof updates.linkedPR === 'number' && + updates.linkedPR > 0 && + updates.linkedPR !== currentLinkedPR ? { ...updates, suppressedGitHubPR: null } : updates } diff --git a/src/shared/worktree/types.ts b/src/shared/worktree/types.ts index 1eb076a08ff..3bf89f8e919 100644 --- a/src/shared/worktree/types.ts +++ b/src/shared/worktree/types.ts @@ -258,6 +258,8 @@ export type DetectedWorktreeListResult = { failureKind?: WorktreeScanFailureKind /** Which catalog this listing describes; additive, older hosts omit it. */ catalogVersion?: WorktreeCatalogVersion + /** The producer's workspaceMetadata position when this listing read metadata; additive, older hosts omit it. */ + workspaceMetadata?: { epoch: string; seq: number } /** The repo's WSL distro is stopped or unverifiable, so rows are stored ones; additive, older hosts omit it. */ wslDistroStopped?: { distro: string; state: 'stopped' | 'unverifiable' } } diff --git a/src/shared/worktree/workspace-metadata-set.ts b/src/shared/worktree/workspace-metadata-set.ts new file mode 100644 index 00000000000..b66b8550e8e --- /dev/null +++ b/src/shared/worktree/workspace-metadata-set.ts @@ -0,0 +1,84 @@ +import { normalizeExecutionHostId, type ExecutionHostId } from '../execution-host' +import { normalizeStoredTaskSourceContext } from '../work-items/task-source-context' +import { normalizeWorkspaceAttachments } from '../workspace-attachments/workspace-attachment-normalization' +import { normalizeWorkspaceLinkedItem } from '../workspace-attachments/workspace-linked-item' +import type { WorktreeMeta } from './meta-types' + +/** A producer's `workspaceMetadata` position: records with `seq` at or below it are in what it read. */ +export type WorkspaceMetadataCursor = { epoch: string; seq: number } + +/** The workspace-metadata fields a producer publishes today; the row widens without changing the envelope. */ +export type WorkspaceMetadataSetRow = Pick< + WorktreeMeta, + 'linkedItems' | 'linkedWorkItem' | 'linkedIssue' | 'linkedTaskSourceContext' +> + +/** + * One addressed `set` record in a producer host's `workspaceMetadata` domain: full current values for one + * workspace, ordered by `seq` within `epoch`. A gap, a new epoch or an unknown row falls back to the listing. + */ +export type WorkspaceMetadataSet = { + kind: 'set' + domain: 'workspaceMetadata' + hostId: ExecutionHostId + epoch: string + seq: number + // Host-qualified worktree identity; repoId and worktreeId route it to the client's listing. + id: string + repoId: string + worktreeId: string + row: WorkspaceMetadataSetRow +} + +export function workspaceMetadataSetRow(meta: WorktreeMeta): WorkspaceMetadataSetRow { + return { + linkedItems: meta.linkedItems, + linkedWorkItem: meta.linkedWorkItem, + linkedIssue: meta.linkedIssue, + linkedTaskSourceContext: meta.linkedTaskSourceContext + } +} + +function isRecord(value: unknown): value is Record { + return typeof value === 'object' && value !== null && !Array.isArray(value) +} + +/** A well-formed record, or null when an older or malformed producer sent something else. */ +export function readWorkspaceMetadataSet(value: unknown): WorkspaceMetadataSet | null { + if (!isRecord(value) || !isRecord(value.row)) { + return null + } + const { kind, domain, epoch, seq, id, repoId, worktreeId, row } = value + const hostId = normalizeExecutionHostId(typeof value.hostId === 'string' ? value.hostId : null) + if ( + kind !== 'set' || + domain !== 'workspaceMetadata' || + !hostId || + typeof epoch !== 'string' || + typeof seq !== 'number' || + !Number.isSafeInteger(seq) || + typeof id !== 'string' || + typeof repoId !== 'string' || + typeof worktreeId !== 'string' || + !Array.isArray(row.linkedItems) || + (row.linkedIssue !== null && typeof row.linkedIssue !== 'number') + ) { + return null + } + return { + kind, + domain, + hostId, + epoch, + seq, + id, + repoId, + worktreeId, + row: { + linkedItems: normalizeWorkspaceAttachments(row.linkedItems), + linkedWorkItem: normalizeWorkspaceLinkedItem(row.linkedWorkItem), + linkedIssue: row.linkedIssue, + linkedTaskSourceContext: normalizeStoredTaskSourceContext(row.linkedTaskSourceContext) + } + } +} From 06f97dee35f42cd528e383ed3d02ba651747b26c Mon Sep 17 00:00:00 2001 From: "buf0-bot[bot]" <252831055+buf0-bot[bot]@users.noreply.github.com> Date: Wed, 11 Nov 2026 00:39:37 +0000 Subject: [PATCH 06/21] Update README downloads badge (#40940) Co-authored-by: buf0-bot[bot] <252831055+buf0-bot[bot]@users.noreply.github.com> --- docs/assets/readme-downloads.svg | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/docs/assets/readme-downloads.svg b/docs/assets/readme-downloads.svg index 25b9a04169b..2892989fee1 100644 --- a/docs/assets/readme-downloads.svg +++ b/docs/assets/readme-downloads.svg @@ -1,5 +1,5 @@ - - downloads: 168m + + downloads: 169m @@ -15,7 +15,7 @@ downloads downloads - 168m - 168m + 169m + 169m From a7a21c8658ecc0ece24a0ce38d58024a5ae9e5df Mon Sep 17 00:00:00 2001 From: Jinjing <6427696+AmethystLiang@users.noreply.github.com> Date: Tue, 10 Nov 2026 16:40:17 -0800 Subject: [PATCH 07/21] Source Control menu: one line per row with a short status (#40919) * Show Source Control menu rows on one line with a short status * Keep Push and Force Push clickable; force push shows remote-only commits it replaces * Match Force Push description to remote-only commits; list review blockers exhaustively * Show short statuses for every blocked review action --- mobile/src/components/ActionSheetModal.tsx | 3 + .../src/components/nav-bar/PullDownMenu.tsx | 16 +++- mobile/src/i18n/locales/en.json | 17 +++- .../mobile-commit-dropdown-copy.ts | 14 +++- .../mobile-remote-dropdown-copy.ts | 11 ++- .../mobile-source-control-actions.ts | 4 +- .../source-control/commit/commit-area.tsx | 63 +++++++------- .../panel/source-control-dropdown-items.ts | 33 +++++++- ...-control-dropdown-remote-localized-copy.ts | 36 +++++++- .../source-control-dropdown-review-items.ts | 73 ++++++++++++++++ src/renderer/src/i18n/locales/en.json | 23 ++++- .../source-control-dropdown-commit-items.ts | 43 +++++++++- .../source-control-dropdown-item-types.ts | 5 ++ .../source-control-dropdown-labels.ts | 15 ++-- .../source-control-dropdown-remote-copy.ts | 18 +++- .../source-control-dropdown-remote-items.ts | 71 ++++++++++++++-- .../source-control-dropdown-status.test.ts | 83 +++++++++++++++++++ 17 files changed, 470 insertions(+), 58 deletions(-) create mode 100644 src/shared/source-control/dropdown/source-control-dropdown-status.test.ts diff --git a/mobile/src/components/ActionSheetModal.tsx b/mobile/src/components/ActionSheetModal.tsx index 774adcc3c40..49a25ebbc78 100644 --- a/mobile/src/components/ActionSheetModal.tsx +++ b/mobile/src/components/ActionSheetModal.tsx @@ -15,6 +15,8 @@ export type ActionSheetAction = { destructive?: boolean disabled?: boolean hint?: string + /** Short trailing state, e.g. "↑2" or "Up to date". */ + status?: string /** Marks the choice that is current, as a trailing check. */ checked?: boolean loading?: boolean @@ -77,6 +79,7 @@ export function ActionSheetContent({ title, message, children, actions, onClose key={action.id ?? action.label} title={action.label} subtitle={action.hint} + value={action.status} leading={ action.renderIcon?.() ?? ( press(action)} accessibilityRole="menuitem" - accessibilityLabel={action.label} + accessibilityLabel={ + action.status ? `${action.label}, ${action.status}` : action.label + } accessibilityHint={action.hint} accessibilityState={ action.checked === undefined ? undefined : { checked: action.checked } @@ -213,6 +215,11 @@ export function PullDownMenu({ visible, anchorRef, groups, onClose }: PullDownMe ) : null} + {action.status ? ( + + {action.status} + + ) : null} {action.checked ? ( ) : action.renderIcon ? ( @@ -280,5 +287,12 @@ const styles = themedStyleSheet(() => ({ hint: { color: colors.textSecondary, fontSize: typography.metaSize + }, + status: { + marginLeft: 'auto', + flexShrink: 0, + color: colors.textSecondary, + fontSize: typography.metaSize, + fontVariant: ['tabular-nums'] } })) diff --git a/mobile/src/i18n/locales/en.json b/mobile/src/i18n/locales/en.json index cac5f2aa90f..ccb6044a3f6 100644 --- a/mobile/src/i18n/locales/en.json +++ b/mobile/src/i18n/locales/en.json @@ -3484,7 +3484,12 @@ "commitThenPullPush": "Commit, then pull and push", "resolveConflicts": "Resolve conflicts before committing", "stageFile": "Stage at least one file to commit", - "enterMessage": "Enter a commit message to commit" + "enterMessage": "Enter a commit message to commit", + "statusConflicts": "Conflicts", + "statusNothingStaged": "Nothing staged", + "statusNeedsMessage": "Needs message", + "statusNotPublished": "Not published", + "statusNoBranch": "No branch" }, "createPrAction": { "createReview": "Create {{titleLabel}}", @@ -3561,7 +3566,15 @@ "nothingToPushTo": "Nothing to push to {{upstreamName}}", "nothingToForcePushTo": "Nothing to force push to {{upstreamName}}", "rebaseFromNamedBase": "Rebase from {{baseRef}}", - "rebaseWithLatestFrom": "Rebase current branch with latest commits from {{baseRef}}" + "rebaseWithLatestFrom": "Rebase current branch with latest commits from {{baseRef}}", + "statusUpToDate": "Up to date", + "statusSetsUpstream": "Sets upstream", + "statusNotPublished": "Not published", + "statusNoBranch": "No branch", + "statusMayFail": "May fail", + "statusPublished": "Published", + "statusStashFirst": "Stash first", + "statusReplaces": "Replaces ↓{{behind}}" }, "actions": { "createPr": "Create PR", diff --git a/mobile/src/source-control/mobile-commit-dropdown-copy.ts b/mobile/src/source-control/mobile-commit-dropdown-copy.ts index 6ed43bd8647..7cc182bde00 100644 --- a/mobile/src/source-control/mobile-commit-dropdown-copy.ts +++ b/mobile/src/source-control/mobile-commit-dropdown-copy.ts @@ -69,6 +69,18 @@ export function mobileCommitDropdownCopy(t: Translate): CommitDropdownCopy { case 'Enter a commit message to commit': return t('sourceControl.commitDropdown.enterMessage', 'Enter a commit message to commit') } - } + }, + commitDisabledStatus: (reason) => { + switch (reason) { + case 'Resolve conflicts before committing': + return t('sourceControl.commitDropdown.statusConflicts', 'Conflicts') + case 'Stage at least one file to commit': + return t('sourceControl.commitDropdown.statusNothingStaged', 'Nothing staged') + case 'Enter a commit message to commit': + return t('sourceControl.commitDropdown.statusNeedsMessage', 'Needs message') + } + }, + statusNotPublished: t('sourceControl.commitDropdown.statusNotPublished', 'Not published'), + statusNoBranch: t('sourceControl.commitDropdown.statusNoBranch', 'No branch') } } diff --git a/mobile/src/source-control/mobile-remote-dropdown-copy.ts b/mobile/src/source-control/mobile-remote-dropdown-copy.ts index a8cea9cbbb3..8f5e9a55ecf 100644 --- a/mobile/src/source-control/mobile-remote-dropdown-copy.ts +++ b/mobile/src/source-control/mobile-remote-dropdown-copy.ts @@ -160,6 +160,15 @@ export function mobileRemoteDropdownCopy(t: Translate): RemoteDropdownCopy { manualForcePush: (ahead, behind, upstreamName) => localizedManualForcePush(ahead, behind, upstreamName, t), unpublishedForcePush: (branchCommitsAhead) => - localizedUnpublishedForcePush(branchCommitsAhead, t) + localizedUnpublishedForcePush(branchCommitsAhead, t), + statusUpToDate: t('sourceControl.remoteDropdown.statusUpToDate', 'Up to date'), + statusSetsUpstream: t('sourceControl.remoteDropdown.statusSetsUpstream', 'Sets upstream'), + statusNotPublished: t('sourceControl.remoteDropdown.statusNotPublished', 'Not published'), + statusNoBranch: t('sourceControl.remoteDropdown.statusNoBranch', 'No branch'), + statusMayFail: t('sourceControl.remoteDropdown.statusMayFail', 'May fail'), + statusPublished: t('sourceControl.remoteDropdown.statusPublished', 'Published'), + statusStashFirst: t('sourceControl.remoteDropdown.statusStashFirst', 'Stash first'), + statusReplaces: (behind) => + t('sourceControl.remoteDropdown.statusReplaces', 'Replaces ↓{{behind}}', { behind }) } } diff --git a/mobile/src/source-control/mobile-source-control-actions.ts b/mobile/src/source-control/mobile-source-control-actions.ts index cd819d69476..c4c453ddea4 100644 --- a/mobile/src/source-control/mobile-source-control-actions.ts +++ b/mobile/src/source-control/mobile-source-control-actions.ts @@ -31,6 +31,7 @@ export type MobileSourceControlAction = { iconKey: MobileSourceControlActionIcon disabled?: boolean hint?: string + status?: string loading?: boolean skipAutoClose?: boolean closeBeforePress?: boolean @@ -84,7 +85,7 @@ export function buildMobileSourceControlActions( const busy = args.busyAction !== null const blockForce = (item: DropdownItem): DropdownItem => args.forcePushBlockedReason - ? { ...item, disabled: true, title: args.forcePushBlockedReason } + ? { ...item, disabled: true, title: args.forcePushBlockedReason, status: undefined } : item const row = ( item: DropdownItem, @@ -98,6 +99,7 @@ export function buildMobileSourceControlActions( disabled: busy || item.disabled, // Mobile has no tooltips: show why a row is disabled, and always what Force Push replaces. hint: item.disabled || item.kind === 'force_push' ? item.title : undefined, + status: item.status, loading: args.busyAction === actionId, skipAutoClose: true, onPress diff --git a/src/renderer/src/components/right-sidebar/source-control/commit/commit-area.tsx b/src/renderer/src/components/right-sidebar/source-control/commit/commit-area.tsx index 4dc129b24b9..bf7a0537a43 100644 --- a/src/renderer/src/components/right-sidebar/source-control/commit/commit-area.tsx +++ b/src/renderer/src/components/right-sidebar/source-control/commit/commit-area.tsx @@ -1,4 +1,4 @@ -import React, { useId, useMemo } from 'react' +import React, { useMemo } from 'react' import { ArrowDownUp, ArrowUp, Check, CloudUpload, GitPullRequestArrow, Plus } from 'lucide-react' import { DropdownMenuContent, @@ -6,6 +6,7 @@ import { DropdownMenuSeparator } from '@/components/ui/dropdown-menu' import { useTranslate } from '@/i18n/use-translate' +import { cn } from '@/lib/class-names' import { hasExpandedCommitFailureDetails, summarizeCommitFailure @@ -80,7 +81,6 @@ export function CommitArea({ onDropdownAction }: CommitAreaProps): React.JSX.Element { const t = useTranslate() - const actionDescriptionId = useId() // Cap rows so long messages scroll internally instead of pushing Commit offscreen. const rows = getCommitMessageTextareaRows(commitMessage) // Spin the primary only for its matching operation; Fetch uses the chevron and commits use isCommitting. @@ -173,41 +173,44 @@ export function CommitArea({ 'More commit and remote actions' ) const dropdownMenuContent = ( - + {dropdownItems.map((entry) => entry.kind === 'separator' ? ( ) : ( -
- { - if (entry.disabled) { - event.preventDefault() - return - } - onDropdownAction(entry.kind) - }} - > - - {entry.label} - {!entry.disabled && ( - {entry.title} + { + if (entry.disabled) { + event.preventDefault() + return + } + onDropdownAction(entry.kind) + }} + > + + {entry.label} + + {entry.status && ( + - - {entry.disabled && ( -

- {entry.title} -

+ {entry.status} +
)} -
+ ) )} diff --git a/src/renderer/src/components/right-sidebar/source-control/panel/source-control-dropdown-items.ts b/src/renderer/src/components/right-sidebar/source-control/panel/source-control-dropdown-items.ts index 00e369e0a99..096ce79280f 100644 --- a/src/renderer/src/components/right-sidebar/source-control/panel/source-control-dropdown-items.ts +++ b/src/renderer/src/components/right-sidebar/source-control/panel/source-control-dropdown-items.ts @@ -39,6 +39,26 @@ function localizedCommitDisabledReason(reason: CommitDisabledReason): string { } } +function localizedCommitDisabledStatus(reason: CommitDisabledReason): string { + switch (reason) { + case 'Resolve conflicts before committing': + return translate( + 'auto.components.right.sidebar.source.control.dropdown.items.statusConflicts', + 'Conflicts' + ) + case 'Stage at least one file to commit': + return translate( + 'auto.components.right.sidebar.source.control.dropdown.items.statusNothingStaged', + 'Nothing staged' + ) + case COMMIT_MESSAGE_REQUIRED_REASON: + return translate( + 'auto.components.right.sidebar.source.control.dropdown.items.statusNeedsMessage', + 'Needs message' + ) + } +} + function localizedCommitCopy(t: Translate = translate): CommitDropdownCopy { return { commit: t('auto.components.right.sidebar.source.control.dropdown.items.2b8e6595fd', 'Commit'), @@ -110,7 +130,16 @@ function localizedCommitCopy(t: Translate = translate): CommitDropdownCopy { 'auto.components.right.sidebar.source.control.dropdown.items.titleCommitThenPullPush', 'Commit, then pull and push' ), - commitDisabledReason: localizedCommitDisabledReason + commitDisabledReason: localizedCommitDisabledReason, + commitDisabledStatus: localizedCommitDisabledStatus, + statusNotPublished: t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusNotPublished', + 'Not published' + ), + statusNoBranch: t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusNoBranch', + 'No branch' + ) } } @@ -183,6 +212,8 @@ export function resolveDropdownItems( 'auto.components.right.sidebar.source.control.dropdown.items.7aad2c0240', 'Hosted review operation in progress…' ), + status: undefined, + statusTone: undefined, disabled: true } ) diff --git a/src/renderer/src/components/right-sidebar/source-control/panel/source-control-dropdown-remote-localized-copy.ts b/src/renderer/src/components/right-sidebar/source-control/panel/source-control-dropdown-remote-localized-copy.ts index 9228b13d6b7..2046a8a9eb2 100644 --- a/src/renderer/src/components/right-sidebar/source-control/panel/source-control-dropdown-remote-localized-copy.ts +++ b/src/renderer/src/components/right-sidebar/source-control/panel/source-control-dropdown-remote-localized-copy.ts @@ -203,6 +203,40 @@ export function localizedRemoteCopy(t: Translate = translate): RemoteDropdownCop syncCounts: localizedSyncCounts, forcePushWithLease: localizedForcePushWithLease, manualForcePush: localizedManualForcePush, - unpublishedForcePush: localizedUnpublishedForcePush + unpublishedForcePush: localizedUnpublishedForcePush, + statusUpToDate: t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusUpToDate', + 'Up to date' + ), + statusSetsUpstream: t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusSetsUpstream', + 'Sets upstream' + ), + statusNotPublished: t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusNotPublished', + 'Not published' + ), + statusNoBranch: t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusNoBranch', + 'No branch' + ), + statusMayFail: t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusMayFail', + 'May fail' + ), + statusPublished: t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusPublished', + 'Published' + ), + statusStashFirst: t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusStashFirst', + 'Stash first' + ), + statusReplaces: (behind) => + t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusReplaces', + 'Replaces ↓{{behind}}', + { behind } + ) } } diff --git a/src/renderer/src/components/right-sidebar/source-control/review/source-control-dropdown-review-items.ts b/src/renderer/src/components/right-sidebar/source-control/review/source-control-dropdown-review-items.ts index 77dba32b803..8bce6a0a62e 100644 --- a/src/renderer/src/components/right-sidebar/source-control/review/source-control-dropdown-review-items.ts +++ b/src/renderer/src/components/right-sidebar/source-control/review/source-control-dropdown-review-items.ts @@ -101,6 +101,77 @@ export function buildHostedReviewDropdownItems( } })() + const createBlockedStatus = ((): string | undefined => { + switch (hostedReviewCreation?.blockedReason) { + case 'dirty': + return t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusCommitFirst', + 'Commit first' + ) + case 'detached_head': + return t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusNoBranch', + 'No branch' + ) + case 'existing_review': + return t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusReviewOpen', + '{{shortLabel}} open', + { + shortLabel: createReviewCopy.shortLabel + } + ) + case 'no_upstream': + return t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusNotPublished', + 'Not published' + ) + case 'auth_required': + return t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusSignIn', + 'Sign in' + ) + case 'base_not_on_remote': + return t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusBaseNotOnRemote', + 'Base not on remote' + ) + case 'default_branch': + return t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusDefaultBranch', + 'Default branch' + ) + case 'fork_head_unsupported': + return t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusForkUnsupported', + 'Fork unsupported' + ) + case 'needs_push': + return t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusPushFirst', + 'Push first' + ) + case 'needs_sync': + return shouldForcePushWithLease + ? t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusForcePushFirst', + 'Force push first' + ) + : t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusSyncFirst', + 'Sync first' + ) + case 'unsupported_provider': + return t( + 'auto.components.right.sidebar.source.control.dropdown.items.statusUnsupported', + 'Unsupported' + ) + case null: + case undefined: + return undefined + } + })() + const createPR: DropdownItem = { kind: 'create_pr', label: t( @@ -116,6 +187,7 @@ export function buildHostedReviewDropdownItems( ) : createBlockedHint, hint: hostedReviewCreation?.canCreate ? undefined : createBlockedHint, + status: hostedReviewCreation?.canCreate ? undefined : createBlockedStatus, disabled: globalBusy || !supportsHostedReviewCreation(hostedReviewCreation?.provider) || @@ -160,6 +232,7 @@ export function buildHostedReviewDropdownItems( ) : createBlockedHint, hint: canPushAndCreate ? undefined : createBlockedHint, + status: canPushAndCreate ? undefined : createBlockedStatus, disabled: !canPushAndCreate } diff --git a/src/renderer/src/i18n/locales/en.json b/src/renderer/src/i18n/locales/en.json index c224064e058..f30ad7b32f9 100644 --- a/src/renderer/src/i18n/locales/en.json +++ b/src/renderer/src/i18n/locales/en.json @@ -14969,7 +14969,28 @@ "titleManualForcePush_other": "Force push {{count}} local commits with lease to update {{upstreamName}}.", "titleUnpublishedForcePushBranch": "Force push this branch with lease and set an upstream if needed.", "titleUnpublishedForcePushCount_one": "Force push {{count}} branch commit with lease and set an upstream if needed.", - "titleUnpublishedForcePushCount_other": "Force push {{count}} branch commits with lease and set an upstream if needed." + "titleUnpublishedForcePushCount_other": "Force push {{count}} branch commits with lease and set an upstream if needed.", + "statusConflicts": "Conflicts", + "statusNothingStaged": "Nothing staged", + "statusNeedsMessage": "Needs message", + "statusNotPublished": "Not published", + "statusNoBranch": "No branch", + "statusUpToDate": "Up to date", + "statusSetsUpstream": "Sets upstream", + "statusMayFail": "May fail", + "statusPublished": "Published", + "statusStashFirst": "Stash first", + "statusReplaces": "Replaces ↓{{behind}}", + "statusCommitFirst": "Commit first", + "statusReviewOpen": "{{shortLabel}} open", + "statusSignIn": "Sign in", + "statusBaseNotOnRemote": "Base not on remote", + "statusDefaultBranch": "Default branch", + "statusForkUnsupported": "Fork unsupported", + "statusPushFirst": "Push first", + "statusForcePushFirst": "Force push first", + "statusSyncFirst": "Sync first", + "statusUnsupported": "Unsupported" } }, "primary": { diff --git a/src/shared/source-control/dropdown/source-control-dropdown-commit-items.ts b/src/shared/source-control/dropdown/source-control-dropdown-commit-items.ts index a3a8e50ba14..45e091a1fb9 100644 --- a/src/shared/source-control/dropdown/source-control-dropdown-commit-items.ts +++ b/src/shared/source-control/dropdown/source-control-dropdown-commit-items.ts @@ -1,6 +1,9 @@ // Share commit blockers across Commit, Commit & Push, and Commit & Sync so disabled reasons agree. -import type { CommitDisabledReason } from '../decisions/source-control-commit-eligibility' +import { + COMMIT_MESSAGE_REQUIRED_REASON, + type CommitDisabledReason +} from '../decisions/source-control-commit-eligibility' import type { DropdownItem } from './source-control-dropdown-item-types' import type { DropdownActionContext } from './source-control-dropdown-action-context' @@ -26,6 +29,20 @@ export type CommitDropdownCopy = { commitThenPullPush: string // Eligibility compares the English reasons, so copy translates them only for display. commitDisabledReason: (reason: CommitDisabledReason) => string + commitDisabledStatus: (reason: CommitDisabledReason) => string + statusNotPublished: string + statusNoBranch: string +} + +export function englishCommitDisabledStatus(reason: CommitDisabledReason): string { + switch (reason) { + case 'Resolve conflicts before committing': + return 'Conflicts' + case 'Stage at least one file to commit': + return 'Nothing staged' + case COMMIT_MESSAGE_REQUIRED_REASON: + return 'Needs message' + } } export const ENGLISH_COMMIT_DROPDOWN_COPY: CommitDropdownCopy = { @@ -47,7 +64,10 @@ export const ENGLISH_COMMIT_DROPDOWN_COPY: CommitDropdownCopy = { publishBranchFirstToSync: 'Publish the branch first to sync commits', useCommitAndForcePush: 'Use Commit & Force Push — remote only has older copies of local commits', commitThenPullPush: 'Commit, then pull and push', - commitDisabledReason: (reason) => reason + commitDisabledReason: (reason) => reason, + commitDisabledStatus: englishCommitDisabledStatus, + statusNotPublished: 'Not published', + statusNoBranch: 'No branch' } export type CommitDropdownItems = { @@ -77,11 +97,20 @@ export function buildCommitDropdownItems( } = ctx const commitDisabledTitle = commitDisabledReason === null ? null : copy.commitDisabledReason(commitDisabledReason) + const commitDisabledStatus = + commitDisabledReason === null ? undefined : copy.commitDisabledStatus(commitDisabledReason) + // Branch blockers outrank commit blockers, matching the title ladders below. + const branchBlockedStatus = publishBlockedByDetachedHead + ? copy.statusNoBranch + : !upstreamLoading && !hasUpstream + ? copy.statusNotPublished + : undefined const commit: DropdownItem = { kind: 'commit', label: copy.commit, title: commitDisabledTitle ?? copy.commitStagedChanges, + status: commitDisabledStatus, disabled: !canCommit } @@ -108,6 +137,12 @@ export function buildCommitDropdownItems( kind: 'commit_push', label: shouldForcePushWithLease ? copy.commitForcePush : copy.commitPush, title: commitPushTitle, + status: + upstreamLoading || publishBlockedByPRLoading || publishBlockedByMergedPR + ? undefined + : !hasUpstream && hasOpenHostedReview && canPushLinkedReviewWithoutUpstream + ? commitDisabledStatus + : (branchBlockedStatus ?? commitDisabledStatus), // Match Push: without upstream, Commit & Push requires a known open-review head. disabled: globalBusy || @@ -145,6 +180,10 @@ export function buildCommitDropdownItems( kind: 'commit_sync', label: copy.commitSync, title: commitSyncTitle, + status: + upstreamLoading || publishBlockedByPRLoading || publishBlockedByMergedPR + ? undefined + : (branchBlockedStatus ?? commitDisabledStatus), disabled: globalBusy || upstreamLoading || diff --git a/src/shared/source-control/dropdown/source-control-dropdown-item-types.ts b/src/shared/source-control/dropdown/source-control-dropdown-item-types.ts index ac0a2b288a5..896f0c391c2 100644 --- a/src/shared/source-control/dropdown/source-control-dropdown-item-types.ts +++ b/src/shared/source-control/dropdown/source-control-dropdown-item-types.ts @@ -32,6 +32,11 @@ export type DropdownItem = { title: string disabled: boolean hint?: string + // Short right-aligned state; `title` keeps the full sentence for screen readers. + status?: string + statusTone?: 'active' | 'warning' + // Nothing to do right now, but still clickable (counts can be stale until a fetch). + idle?: boolean variant?: 'default' | 'destructive' } diff --git a/src/shared/source-control/dropdown/source-control-dropdown-labels.ts b/src/shared/source-control/dropdown/source-control-dropdown-labels.ts index a24c58719bf..4df2193f00a 100644 --- a/src/shared/source-control/dropdown/source-control-dropdown-labels.ts +++ b/src/shared/source-control/dropdown/source-control-dropdown-labels.ts @@ -14,15 +14,16 @@ export function describeSyncCounts(ahead: number, behind: number): string { return `Pull ${behind}, push ${ahead}` } -export function formatCountLabel(base: string, count: number): string { - return count > 0 ? `${base} (${count})` : base +export function formatAheadStatus(ahead: number): string { + return `↑${ahead}` } -export function formatSyncLabel(base: string, ahead: number, behind: number): string { - if (ahead === 0 && behind === 0) { - return base - } - return `${base} (↓${behind} ↑${ahead})` +export function formatBehindStatus(behind: number): string { + return `↓${behind}` +} + +export function formatSyncStatus(ahead: number, behind: number): string { + return `↓${behind} ↑${ahead}` } export function formatForcePushTitle( diff --git a/src/shared/source-control/dropdown/source-control-dropdown-remote-copy.ts b/src/shared/source-control/dropdown/source-control-dropdown-remote-copy.ts index e8130c32f27..c2160c25104 100644 --- a/src/shared/source-control/dropdown/source-control-dropdown-remote-copy.ts +++ b/src/shared/source-control/dropdown/source-control-dropdown-remote-copy.ts @@ -64,6 +64,14 @@ export type RemoteDropdownCopy = { forcePushWithLease: (branchCommitsAhead: number | undefined, upstreamName?: string) => string manualForcePush: (ahead: number, behind: number, upstreamName?: string) => string unpublishedForcePush: (branchCommitsAhead: number | undefined) => string + statusUpToDate: string + statusSetsUpstream: string + statusNotPublished: string + statusNoBranch: string + statusMayFail: string + statusPublished: string + statusStashFirst: string + statusReplaces: (behind: number) => string } export const ENGLISH_REMOTE_DROPDOWN_COPY: RemoteDropdownCopy = { @@ -121,5 +129,13 @@ export const ENGLISH_REMOTE_DROPDOWN_COPY: RemoteDropdownCopy = { syncCounts: describeSyncCounts, forcePushWithLease: formatForcePushTitle, manualForcePush: formatManualForcePushTitle, - unpublishedForcePush: formatUnpublishedForcePushTitle + unpublishedForcePush: formatUnpublishedForcePushTitle, + statusUpToDate: 'Up to date', + statusSetsUpstream: 'Sets upstream', + statusNotPublished: 'Not published', + statusNoBranch: 'No branch', + statusMayFail: 'May fail', + statusPublished: 'Published', + statusStashFirst: 'Stash first', + statusReplaces: (behind) => `Replaces ↓${behind}` } diff --git a/src/shared/source-control/dropdown/source-control-dropdown-remote-items.ts b/src/shared/source-control/dropdown/source-control-dropdown-remote-items.ts index 997b30d3f65..8219d7d8249 100644 --- a/src/shared/source-control/dropdown/source-control-dropdown-remote-items.ts +++ b/src/shared/source-control/dropdown/source-control-dropdown-remote-items.ts @@ -1,9 +1,10 @@ import type { DropdownItem } from './source-control-dropdown-item-types' import type { DropdownActionContext } from './source-control-dropdown-action-context' import { - formatCountLabel, + formatAheadStatus, + formatBehindStatus, formatRebaseBaseRef, - formatSyncLabel + formatSyncStatus } from './source-control-dropdown-labels' import { ENGLISH_REMOTE_DROPDOWN_COPY, @@ -45,10 +46,32 @@ export function buildRemoteDropdownItems( shouldForcePushWithLease, pushLabelCount } = ctx - + const remoteBlockedStatus = publishBlockedByDetachedHead + ? copy.statusNoBranch + : !upstreamLoading && !hasUpstream + ? copy.statusNotPublished + : undefined + // Pull-side counts can be stale until a fetch, and these actions fetch first, so they stay clickable. + const incomingStatus = (incoming: string | undefined): Partial => { + if (remoteBlockedStatus || upstreamLoading) { + return { status: remoteBlockedStatus } + } + return incoming + ? { status: incoming, statusTone: 'active' } + : { status: copy.statusUpToDate, idle: true } + } const push: DropdownItem = { kind: 'push', - label: formatCountLabel(copy.push, ahead), + label: copy.push, + ...(publishBlockedByDetachedHead + ? { status: copy.statusNoBranch } + : pushBlockedByOpenHostedReviewTarget || upstreamLoading || canPushUntrackedHostedReview + ? {} + : !hasUpstream + ? { status: copy.statusSetsUpstream } + : ahead > 0 + ? { status: formatAheadStatus(ahead), statusTone: 'active' as const } + : { status: copy.statusUpToDate, idle: true }), title: publishBlockedByDetachedHead ? copy.checkOutBranchToPush : pushBlockedByOpenHostedReviewTarget @@ -74,7 +97,16 @@ export function buildRemoteDropdownItems( const forcePush: DropdownItem = { kind: 'force_push', - label: formatCountLabel(copy.forcePush, pushLabelCount), + label: copy.forcePush, + ...(publishBlockedByDetachedHead + ? { status: copy.statusNoBranch } + : pushBlockedByOpenHostedReviewTarget || upstreamLoading || !hasUpstream + ? {} + : behind > 0 && !shouldForcePushWithLease + ? { status: copy.statusReplaces(behind), statusTone: 'warning' as const } + : pushLabelCount > 0 + ? { status: formatAheadStatus(pushLabelCount), statusTone: 'active' as const } + : { status: copy.statusUpToDate, idle: true }), title: publishBlockedByDetachedHead ? copy.checkOutBranchToForcePush : pushBlockedByOpenHostedReviewTarget @@ -83,7 +115,7 @@ export function buildRemoteDropdownItems( ? copy.unpublishedForcePush(branchCommitsAhead) : !hasUpstream ? copy.unpublishedForcePush(branchCommitsAhead) - : pushLabelCount === 0 + : pushLabelCount === 0 && behind === 0 ? upstreamStatus?.upstreamName ? copy.nothingToForcePushTo(upstreamStatus.upstreamName) : copy.nothingToForcePush @@ -96,7 +128,10 @@ export function buildRemoteDropdownItems( const pull: DropdownItem = { kind: 'pull', - label: formatCountLabel(copy.pull, behind), + label: copy.pull, + ...incomingStatus( + behind > 0 && !shouldForcePushWithLease ? formatBehindStatus(behind) : undefined + ), title: upstreamLoading ? copy.checkingBranchStatus : publishBlockedByPRLoading @@ -117,7 +152,16 @@ export function buildRemoteDropdownItems( const fastForward: DropdownItem = { kind: 'fast_forward', - label: formatCountLabel(copy.fastForward, behind), + label: copy.fastForward, + ...(behind > 0 && + ahead > 0 && + !shouldForcePushWithLease && + !remoteBlockedStatus && + !upstreamLoading + ? { status: copy.statusMayFail, statusTone: 'warning' as const } + : incomingStatus( + behind > 0 && !shouldForcePushWithLease ? formatBehindStatus(behind) : undefined + )), title: upstreamLoading ? copy.checkingBranchStatus : publishBlockedByPRLoading @@ -140,7 +184,10 @@ export function buildRemoteDropdownItems( const sync: DropdownItem = { kind: 'sync', - label: formatSyncLabel(copy.sync, ahead, behind), + label: copy.sync, + ...(shouldForcePushWithLease + ? {} + : incomingStatus(ahead > 0 || behind > 0 ? formatSyncStatus(ahead, behind) : undefined)), title: upstreamLoading ? copy.checkingBranchStatus : publishBlockedByPRLoading @@ -178,6 +225,10 @@ export function buildRemoteDropdownItems( } return copy.rebaseWithLatestFrom(rebaseBaseLabel) })(), + status: + rebaseBaseLabel && hasRemoteBaseRef && hasDirtyLocalChanges + ? copy.statusStashFirst + : undefined, disabled: globalBusy || !rebaseBaseRef || !hasRemoteBaseRef } @@ -198,6 +249,8 @@ export function buildRemoteDropdownItems( : publishBlockedByDetachedHead ? copy.publishNoBranch : copy.publishBranch, + // Detached HEAD already reads "No Branch" in the label. + status: hasUpstream && !upstreamLoading ? copy.statusPublished : undefined, title: upstreamLoading ? copy.checkingBranchStatus : publishBlockedByPRLoading diff --git a/src/shared/source-control/dropdown/source-control-dropdown-status.test.ts b/src/shared/source-control/dropdown/source-control-dropdown-status.test.ts new file mode 100644 index 00000000000..0e6b68c38bb --- /dev/null +++ b/src/shared/source-control/dropdown/source-control-dropdown-status.test.ts @@ -0,0 +1,83 @@ +import { describe, expect, it } from 'vitest' +import type { DropdownActionInputs } from './source-control-dropdown-item-types' +import { deriveDropdownActionContext } from './source-control-dropdown-action-context' +import { buildRemoteDropdownItems } from './source-control-dropdown-remote-items' +import { buildCommitDropdownItems } from './source-control-dropdown-commit-items' + +function inputs(overrides: Partial): DropdownActionInputs { + return { + stagedCount: 0, + hasUnstagedChanges: false, + hasStageableChanges: false, + hasPartiallyStagedChanges: false, + hasMessage: false, + hasUnresolvedConflicts: false, + isCommitting: false, + isRemoteOperationActive: false, + upstreamStatus: { hasUpstream: true, upstreamName: 'origin/feature', ahead: 0, behind: 0 }, + hasCurrentBranch: true, + rebaseBaseRef: 'origin/main', + ...overrides + } +} + +function build(overrides: Partial) { + const ctx = deriveDropdownActionContext(inputs(overrides)) + return { ...buildRemoteDropdownItems(ctx), ...buildCommitDropdownItems(ctx) } +} + +describe('dropdown row status', () => { + it('moves counts out of labels and keeps idle rows clickable', () => { + const upToDate = build({}) + // pushRemote can target a fork that tracked-upstream counts do not describe. + expect(upToDate.push).toMatchObject({ label: 'Push', status: 'Up to date', disabled: false }) + expect(upToDate.forcePush).toMatchObject({ status: 'Up to date', disabled: false }) + // Counts can be stale until a fetch, so fetching actions stay clickable. + expect(upToDate.pull).toMatchObject({ status: 'Up to date', idle: true, disabled: false }) + expect(upToDate.sync).toMatchObject({ status: 'Up to date', idle: true, disabled: false }) + + const diverged = build({ + upstreamStatus: { hasUpstream: true, upstreamName: 'origin/feature', ahead: 1, behind: 3 } + }) + expect(diverged.push).toMatchObject({ label: 'Push', status: '↑1', disabled: false }) + expect(diverged.forcePush).toMatchObject({ status: 'Replaces ↓3', statusTone: 'warning' }) + expect(diverged.pull).toMatchObject({ label: 'Pull', status: '↓3' }) + expect(diverged.fastForward.status).toBe('May fail') + expect(diverged.sync).toMatchObject({ label: 'Sync', status: '↓3 ↑1' }) + }) + + it('lets force push roll back remote-only commits when nothing is ahead', () => { + const items = build({ + upstreamStatus: { hasUpstream: true, upstreamName: 'origin/feature', ahead: 0, behind: 2 } + }) + expect(items.forcePush).toMatchObject({ + status: 'Replaces ↓2', + statusTone: 'warning', + disabled: false + }) + expect(items.push.disabled).toBe(false) + }) + + it('explains unpublished branches with staged files', () => { + const items = build({ + stagedCount: 2, + upstreamStatus: { hasUpstream: false, ahead: 0, behind: 0 } + }) + expect(items.commit.status).toBe('Needs message') + expect(items.commitPush.status).toBe('Not published') + expect(items.push.status).toBe('Sets upstream') + expect(items.pull).toMatchObject({ status: 'Not published', disabled: true }) + expect(items.rebase.status).toBe('Stash first') + }) + + it('reports detached HEAD as no branch', () => { + const items = build({ + hasCurrentBranch: false, + upstreamStatus: { hasUpstream: false, ahead: 0, behind: 0 } + }) + expect(items.commit.status).toBe('Nothing staged') + expect(items.commitSync.status).toBe('No branch') + expect(items.push).toMatchObject({ status: 'No branch', disabled: true }) + expect(items.sync.status).toBe('No branch') + }) +}) From c58ffc004228057d6bdaa2176e4c390f56c04171 Mon Sep 17 00:00:00 2001 From: Kelvin Amoaba <97001695+AmoabaKelvin@users.noreply.github.com> Date: Wed, 11 Nov 2026 00:41:08 +0000 Subject: [PATCH 08/21] revert(mobile): restore the scrolling session tab strip (#40938) Reverts #40592 (12e809c7556). The session bar goes back to the scrolling strip of tab cells, the + menu goes back to the action sheet, and the header subtitle shows the tab count again. --- mobile/src/agents/agent-list-row-text.ts | 11 +- mobile/src/components/SwipeRowActions.tsx | 9 +- mobile/src/i18n/locales/de.json | 3 +- mobile/src/i18n/locales/en.json | 29 +- mobile/src/i18n/locales/es.json | 3 +- mobile/src/i18n/locales/fr.json | 3 +- mobile/src/i18n/locales/ja.json | 3 +- mobile/src/i18n/locales/ko.json | 3 +- mobile/src/i18n/locales/pt.json | 3 +- mobile/src/i18n/locales/ru.json | 3 +- mobile/src/i18n/locales/tr.json | 3 +- mobile/src/i18n/locales/zh-Hant.json | 3 +- mobile/src/i18n/locales/zh.json | 3 +- mobile/src/session/MobileSessionHeader.tsx | 6 +- mobile/src/session/MobileSessionSheets.tsx | 84 ++-- .../session/mobile-session-frame-styles.ts | 102 +++++ mobile/src/session/tabs/MobileNewTabSheet.tsx | 183 --------- .../src/session/tabs/MobileSessionTabBar.tsx | 361 ------------------ .../src/session/tabs/MobileSessionTabMark.tsx | 28 -- .../session/tabs/MobileSessionTabStrip.tsx | 327 ++++++++++++++++ .../session/tabs/MobileSessionTabsSheet.tsx | 220 ----------- .../session/tabs/mobile-session-tab-rows.ts | 121 ------ mobile/src/session/tabs/tab-strip-scroll.ts | 97 +++++ .../session/use-mobile-session-foundation.ts | 2 - ...use-mobile-session-panel-route-actions.tsx | 176 +++++---- .../use-mobile-session-presentation.ts | 7 +- .../use-mobile-session-screen-state.ts | 12 +- .../use-mobile-session-keyboard-state.ts | 67 +++- mobile/src/worktree/agent-row-display.ts | 2 +- .../session-worktree-record-reply-schema.ts | 8 +- mobile/src/worktree/use-live-worktree-name.ts | 19 +- 31 files changed, 810 insertions(+), 1091 deletions(-) delete mode 100644 mobile/src/session/tabs/MobileNewTabSheet.tsx delete mode 100644 mobile/src/session/tabs/MobileSessionTabBar.tsx delete mode 100644 mobile/src/session/tabs/MobileSessionTabMark.tsx create mode 100644 mobile/src/session/tabs/MobileSessionTabStrip.tsx delete mode 100644 mobile/src/session/tabs/MobileSessionTabsSheet.tsx delete mode 100644 mobile/src/session/tabs/mobile-session-tab-rows.ts diff --git a/mobile/src/agents/agent-list-row-text.ts b/mobile/src/agents/agent-list-row-text.ts index 194a77e7616..c3e9d8ef8b7 100644 --- a/mobile/src/agents/agent-list-row-text.ts +++ b/mobile/src/agents/agent-list-row-text.ts @@ -21,14 +21,7 @@ function oneLine(text: string | null | undefined): string { return text?.replace(/\s+/g, ' ').trim() ?? '' } -// A worktree agent row or a session tab's AgentStatusEntry, which leaves these optional. -type AgentPreviewSource = Pick & { - lastAssistantMessage?: string | null - toolName?: string | null - toolInput?: string | null -} - -function lastMessage(agent: AgentPreviewSource): string { +function lastMessage(agent: RuntimeWorktreeAgentRow): string { const message = agent.lastAssistantMessage return message ? oneLine(withoutNativeChatVisualDirectiveLines(message)) : '' } @@ -38,7 +31,7 @@ function lastMessage(agent: AgentPreviewSource): string { * (`group` null) shows no tool, since the tool call is long over. */ export function agentListPreview( - agent: AgentPreviewSource, + agent: RuntimeWorktreeAgentRow, group: AgentStatusGroup | null, title: string ): string { diff --git a/mobile/src/components/SwipeRowActions.tsx b/mobile/src/components/SwipeRowActions.tsx index b070de4ff38..704077f636d 100644 --- a/mobile/src/components/SwipeRowActions.tsx +++ b/mobile/src/components/SwipeRowActions.tsx @@ -84,8 +84,6 @@ export function SwipeRowActions({ }) // Web has no worklet plugin; mapper hooks need this explicit dependency list. const rowStyle = useAnimatedStyle(() => ({ transform: [{ translateX: offset.value }] }), [offset]) - // Hidden at rest, so the action can't show through a rounded group's antialiased corners. - const actionsStyle = useAnimatedStyle(() => ({ opacity: offset.value === 0 ? 0 : 1 }), [offset]) // A tap on an open row closes it rather than pressing the row. const guardPress = (onPress: () => void) => () => { // The click a drag release produces belongs to the drag, not to the row. @@ -126,10 +124,8 @@ export function SwipeRowActions({ } return ( - - {leading ? renderAction(leading, 'leading') : null} - {trailing ? renderAction(trailing, 'trailing') : null} - + {leading ? renderAction(leading, 'leading') : null} + {trailing ? renderAction(trailing, 'trailing') : null} {/* pan-y keeps the page scrolling vertically over the row on the web. */} {children(guardPress)} @@ -148,7 +144,6 @@ const styles = themedStyleSheet(() => ({ justifyContent: 'center', gap: 2 }, - actions: { position: 'absolute', top: 0, right: 0, bottom: 0, left: 0 }, leading: { left: 0 }, trailing: { right: 0 }, actionPressed: { opacity: feedback.pressedOpacity }, diff --git a/mobile/src/i18n/locales/de.json b/mobile/src/i18n/locales/de.json index c8c64c2cf93..4ace38e5fa1 100644 --- a/mobile/src/i18n/locales/de.json +++ b/mobile/src/i18n/locales/de.json @@ -2454,7 +2454,8 @@ "tabStrip": { "checkingCapabilities": "Desktop-Funktionen werden geprüft – gleich erneut versuchen", "newTab": "Neuen Tab erstellen", - "quickCommandsUpdateRequired": "Desktop-Update für Schnellbefehle erforderlich" + "quickCommandsUpdateRequired": "Desktop-Update für Schnellbefehle erforderlich", + "tabStatus": "{{title}}, {{status}}" }, "terminalActionSheetActions": { "clearTerminal": "Terminal leeren", diff --git a/mobile/src/i18n/locales/en.json b/mobile/src/i18n/locales/en.json index ccb6044a3f6..47071529cfc 100644 --- a/mobile/src/i18n/locales/en.json +++ b/mobile/src/i18n/locales/en.json @@ -2184,10 +2184,10 @@ "renameTerminal": "Rename Terminal", "renameTerminalFailed": "Couldn't rename the terminal. {{reason}}", "terminalName": "Terminal name", - "allowAutoClose": "Allow Auto-Close", - "agents": "Agents" + "allowAutoClose": "Allow Auto-Close" }, "tabStrip": { + "tabStatus": "{{title}}, {{status}}", "newTab": "New tab", "quickCommandsUpdateRequired": "Desktop update required for quick commands", "checkingCapabilities": "Checking desktop capabilities — try again in a moment" @@ -2889,6 +2889,8 @@ }, "presentation": { "loadingTabs": "Loading tabs", + "oneTab": "1 tab", + "tabCount": "{{tabCount}} tabs", "tapToRetry": "{{status}} — tap to retry" }, "documentReaders": { @@ -2926,29 +2928,6 @@ "skip": "Skip", "stop": "Stop", "yes": "Yes" - }, - "tabsSheet": { - "needsYou": "Needs you", - "working": "Working", - "everythingElse": "Everything else", - "open": "Open", - "closeTab": "Close {{title}}" - }, - "tabRows": { - "browser": "Browser", - "markdown": "Markdown note", - "file": "File", - "chat": "Chat", - "terminal": "Terminal" - }, - "tabBar": { - "oneTab": "1 tab", - "tabCount": "{{count}} tabs", - "anotherNeedsYou": "another tab needs you", - "actionsHint": "Shows actions for this tab", - "next": "Next tab", - "previous": "Previous tab", - "hint": "Shows all tabs" } }, "chat": { diff --git a/mobile/src/i18n/locales/es.json b/mobile/src/i18n/locales/es.json index 78f372f2264..f994d2c966b 100644 --- a/mobile/src/i18n/locales/es.json +++ b/mobile/src/i18n/locales/es.json @@ -2524,7 +2524,8 @@ "tabStrip": { "checkingCapabilities": "Comprobando las funciones del escritorio. Inténtalo de nuevo en un momento.", "newTab": "Nueva pestaña", - "quickCommandsUpdateRequired": "Actualiza el escritorio para usar los comandos rápidos." + "quickCommandsUpdateRequired": "Actualiza el escritorio para usar los comandos rápidos.", + "tabStatus": "{{title}}, {{status}}" }, "terminalActionSheetActions": { "clearTerminal": "Limpiar terminal", diff --git a/mobile/src/i18n/locales/fr.json b/mobile/src/i18n/locales/fr.json index 20d2d3b5828..966264702ec 100644 --- a/mobile/src/i18n/locales/fr.json +++ b/mobile/src/i18n/locales/fr.json @@ -2524,7 +2524,8 @@ "tabStrip": { "checkingCapabilities": "Vérification des fonctionnalités de l'ordinateur — réessayez dans un instant", "newTab": "Nouvel onglet", - "quickCommandsUpdateRequired": "Mettez à jour Orca sur l'ordinateur pour utiliser les commandes rapides" + "quickCommandsUpdateRequired": "Mettez à jour Orca sur l'ordinateur pour utiliser les commandes rapides", + "tabStatus": "{{title}}, {{status}}" }, "terminalActionSheetActions": { "clearTerminal": "Effacer le terminal", diff --git a/mobile/src/i18n/locales/ja.json b/mobile/src/i18n/locales/ja.json index c6ee2e0fa4f..39364412640 100644 --- a/mobile/src/i18n/locales/ja.json +++ b/mobile/src/i18n/locales/ja.json @@ -2450,7 +2450,8 @@ "tabStrip": { "checkingCapabilities": "デスクトップの対応機能を確認中です。少し待ってから再試行してください。", "newTab": "新規タブ", - "quickCommandsUpdateRequired": "クイックコマンドの利用にはデスクトップの更新が必要です。" + "quickCommandsUpdateRequired": "クイックコマンドの利用にはデスクトップの更新が必要です。", + "tabStatus": "{{title}}、{{status}}" }, "terminalActionSheetActions": { "clearTerminal": "ターミナルをクリア", diff --git a/mobile/src/i18n/locales/ko.json b/mobile/src/i18n/locales/ko.json index 2b51f3ce5ae..c8304776585 100644 --- a/mobile/src/i18n/locales/ko.json +++ b/mobile/src/i18n/locales/ko.json @@ -2450,7 +2450,8 @@ "tabStrip": { "checkingCapabilities": "데스크톱에서 지원하는 기능을 확인 중입니다. 잠시 후 다시 시도하세요.", "newTab": "새 탭", - "quickCommandsUpdateRequired": "빠른 명령을 사용하려면 데스크톱 업데이트가 필요합니다." + "quickCommandsUpdateRequired": "빠른 명령을 사용하려면 데스크톱 업데이트가 필요합니다.", + "tabStatus": "{{title}}, {{status}}" }, "terminalActionSheetActions": { "clearTerminal": "터미널 지우기", diff --git a/mobile/src/i18n/locales/pt.json b/mobile/src/i18n/locales/pt.json index 0d4a2150210..c065a24a76f 100644 --- a/mobile/src/i18n/locales/pt.json +++ b/mobile/src/i18n/locales/pt.json @@ -2524,7 +2524,8 @@ "tabStrip": { "checkingCapabilities": "Verificando os recursos do desktop — tente novamente em instantes", "newTab": "Nova aba", - "quickCommandsUpdateRequired": "Atualize o desktop para usar comandos rápidos" + "quickCommandsUpdateRequired": "Atualize o desktop para usar comandos rápidos", + "tabStatus": "{{title}}, {{status}}" }, "terminalActionSheetActions": { "clearTerminal": "Limpar Terminal", diff --git a/mobile/src/i18n/locales/ru.json b/mobile/src/i18n/locales/ru.json index cec7d8ced57..1333beff214 100644 --- a/mobile/src/i18n/locales/ru.json +++ b/mobile/src/i18n/locales/ru.json @@ -2594,7 +2594,8 @@ "tabStrip": { "checkingCapabilities": "Проверяем возможности Orca на компьютере — повторите попытку чуть позже", "newTab": "Новая вкладка", - "quickCommandsUpdateRequired": "Для быстрых команд обновите Orca на компьютере" + "quickCommandsUpdateRequired": "Для быстрых команд обновите Orca на компьютере", + "tabStatus": "{{title}}, {{status}}" }, "terminalActionSheetActions": { "clearTerminal": "Очистить терминал", diff --git a/mobile/src/i18n/locales/tr.json b/mobile/src/i18n/locales/tr.json index 81be23db055..8c58ec1fe2e 100644 --- a/mobile/src/i18n/locales/tr.json +++ b/mobile/src/i18n/locales/tr.json @@ -2454,7 +2454,8 @@ "tabStrip": { "checkingCapabilities": "Masaüstü özellikleri kontrol ediliyor — birazdan tekrar deneyin", "newTab": "Yeni sekme", - "quickCommandsUpdateRequired": "Hızlı komutlar için masaüstü uygulaması güncellenmeli" + "quickCommandsUpdateRequired": "Hızlı komutlar için masaüstü uygulaması güncellenmeli", + "tabStatus": "{{title}}, {{status}}" }, "terminalActionSheetActions": { "clearTerminal": "Terminali Temizle", diff --git a/mobile/src/i18n/locales/zh-Hant.json b/mobile/src/i18n/locales/zh-Hant.json index 535fb5d378a..3e4a1c25efc 100644 --- a/mobile/src/i18n/locales/zh-Hant.json +++ b/mobile/src/i18n/locales/zh-Hant.json @@ -2450,7 +2450,8 @@ "tabStrip": { "checkingCapabilities": "正在檢查桌面版支援的功能,請稍後再試", "newTab": "新分頁", - "quickCommandsUpdateRequired": "需更新桌面版才能使用快速命令" + "quickCommandsUpdateRequired": "需更新桌面版才能使用快速命令", + "tabStatus": "{{title}},{{status}}" }, "terminalActionSheetActions": { "clearTerminal": "清除終端機", diff --git a/mobile/src/i18n/locales/zh.json b/mobile/src/i18n/locales/zh.json index f845b0e146d..820e95b80a3 100644 --- a/mobile/src/i18n/locales/zh.json +++ b/mobile/src/i18n/locales/zh.json @@ -2450,7 +2450,8 @@ "tabStrip": { "checkingCapabilities": "正在检查桌面支持的功能,请稍后重试", "newTab": "新标签页", - "quickCommandsUpdateRequired": "使用快捷命令需更新桌面应用" + "quickCommandsUpdateRequired": "使用快捷命令需更新桌面应用", + "tabStatus": "{{title}},{{status}}" }, "terminalActionSheetActions": { "clearTerminal": "清空终端", diff --git a/mobile/src/session/MobileSessionHeader.tsx b/mobile/src/session/MobileSessionHeader.tsx index 9020e7d7486..c7593258e42 100644 --- a/mobile/src/session/MobileSessionHeader.tsx +++ b/mobile/src/session/MobileSessionHeader.tsx @@ -6,7 +6,7 @@ import { NavBar, NavBarRefreshLine } from '../components/nav-bar/NavBar' import { useRefreshHint } from '../query/refresh-hint' import { NavBarButton } from '../components/nav-bar/NavBarButton' import { MobileSessionHeaderMoreMenu } from './MobileSessionHeaderMoreMenu' -import { MobileSessionTabBar } from './tabs/MobileSessionTabBar' +import { MobileSessionTabStrip } from './tabs/MobileSessionTabStrip' import { MobileSessionTitleBlock } from './MobileSessionTitleBlock' import type { SessionTitleCard } from './tabs/use-session-title-card' import { styles } from './mobile-session-styles' @@ -58,11 +58,11 @@ export function MobileSessionHeader({ // Why: a sideways phone is ~390pt tall, so the tabs share the top bar instead of taking a second row. const hasTabs = visibleTabs.length > 0 const tabsInTopBar = (layout.isLandscapePhone || sideRail) && hasTabs - // Bar buttons otherwise leave the tab bar too narrow to read on a narrow header. + // Bar buttons otherwise leave less than two readable tabs on a narrow header. const actionsCollapsed = tabsInTopBar && (headerWidth ?? layout.width) < 600 const sourceControlInBar = !isFloatingWorkspaceRoute && !actionsCollapsed const tabStrip = hasTabs ? ( - - { + setShowCreateTabDrawer(false) + void handleCreateTerminal() + } + }, + // Why: browser/markdown creation resolve a real worktree on the host + // (browser.tabCreate, files.createFile); the floating sentinel is + // terminal-only over RPC, so those options hide there. + ...(isFloatingWorkspaceRoute + ? [] + : [ + { + id: 'browser', + label: t('session.sheets.browser', 'Browser'), + icon: Globe, + closeBeforePress: true, + onPress: () => { + if (browserScreencastSupported !== true) { + showToast( + t( + 'session.sheets.browserUpdateRequired', + 'Desktop update required for mobile browser streaming' + ), + 1600 + ) + return + } + setShowCreateBrowserModal(true) + } + }, + ...(onPairedHost + ? [ + { + id: 'markdownNote', + label: t('session.sheets.markdownNote', 'Markdown Note'), + icon: FileText, + onPress: () => { + setShowCreateTabDrawer(false) + void handleCreateMarkdownNote() + } + } + ] + : []) + ]) + ]} onClose={() => setShowCreateTabDrawer(false)} - onCreateAgent={(agent) => void handleCreateTerminal(agent)} - onCreateTerminal={() => void handleCreateTerminal()} - onCreateBrowser={() => { - if (browserScreencastSupported !== true) { - showToast( - t( - 'session.sheets.browserUpdateRequired', - 'Desktop update required for mobile browser streaming' - ), - 1600 - ) - return - } - setShowCreateBrowserModal(true) - }} - onCreateMarkdownNote={() => void handleCreateMarkdownNote()} /> ({ color: colors.textSecondary, fontSize: typography.metaSize }, + tabBar: { + flexDirection: 'row', + alignItems: 'center', + borderTopWidth: 1, + borderTopColor: colors.borderSubtle + }, + // Cancels the top bar's vertical padding so tab borders run the bar's full height. + tabBarInTopBar: { + flex: 1, + minWidth: 0, + flexDirection: 'row', + alignItems: 'center', + alignSelf: 'stretch', + marginVertical: -spacing.xs, + marginLeft: spacing.md, + borderLeftWidth: 1, + borderLeftColor: colors.borderSubtle + }, + tabScrollFrame: { + flex: 1, + minWidth: 0, + flexDirection: 'row', + alignItems: 'center', + alignSelf: 'stretch' + }, + tabEdgeHint: { + position: 'absolute', + top: 0, + bottom: 0, + justifyContent: 'center', + paddingHorizontal: spacing.xs, + backgroundColor: colors.bgPanel + }, + tabEdgeHintLeft: { + left: 0, + borderRightWidth: 1, + borderRightColor: colors.borderSubtle + }, + tabEdgeHintRight: { + right: 0, + borderLeftWidth: 1, + borderLeftColor: colors.borderSubtle + }, + // Opaque like the chrome behind it: see the chat page scroll (#36553). + tabScroll: { + flex: 1, + maxHeight: 44, + backgroundColor: colors.bgPanel + }, + tabScrollInTopBar: { + alignSelf: 'stretch', + maxHeight: 44 + }, + tabContent: { + paddingLeft: spacing.md, + paddingRight: spacing.sm + }, + tabContentInTopBar: { + paddingLeft: 0 + }, + tab: { + minWidth: 44, + maxWidth: 128, + minHeight: 44, + alignItems: 'center', + justifyContent: 'center', + paddingHorizontal: spacing.sm, + paddingVertical: spacing.sm, + borderBottomWidth: 2, + borderBottomColor: 'transparent', + borderRightWidth: 1, + borderRightColor: colors.borderSubtle + }, + tabFirst: { + borderLeftWidth: 1, + borderLeftColor: colors.borderSubtle + }, + tabActive: { + borderBottomColor: colors.textSecondary, + backgroundColor: colors.bgRaised + }, + tabPressed: { + backgroundColor: colors.translucentFill + }, + tabLabelRow: { + maxWidth: '100%', + flexDirection: 'row', + alignItems: 'center', + gap: spacing.xs + }, + // Prevent web SVG shrink before the title ellipsizes. + tabIcon: { + flexShrink: 0 + }, + tabText: { + flexShrink: 1, + color: colors.textSecondary, + fontSize: 13 + }, + tabTextActive: { + color: colors.textPrimary + }, newTerminalButtonPressed: { backgroundColor: colors.bgRaised }, diff --git a/mobile/src/session/tabs/MobileNewTabSheet.tsx b/mobile/src/session/tabs/MobileNewTabSheet.tsx deleted file mode 100644 index 0f7730ddd75..00000000000 --- a/mobile/src/session/tabs/MobileNewTabSheet.tsx +++ /dev/null @@ -1,183 +0,0 @@ -import { useRef } from 'react' -import { Pressable, Text, View } from 'react-native' -import { FileText, Globe, SquareTerminal } from 'lucide-react-native' -import type { TuiAgent } from '../../../../src/shared/agent-launch/catalog/tui-agent' -import { ActionSheetContent, type ActionSheetAction } from '../../components/ActionSheetModal' -import { BottomDrawer } from '../../components/bottom-drawer/BottomDrawer' -import { - LIST_LEADING_INSET, - ListGroup, - ListRow, - SectionHeader -} from '../../components/list/ListGroup' -import { MobileAgentIcon } from '../../components/MobileAgentIcon' -import { SheetHeader } from '../../components/nav-bar/SheetHeader' -import { useMobileTranslation } from '../../i18n/use-mobile-translation' -import { colors, radii, spacing } from '../../theme/mobile-theme' -import { themedStyleSheet } from '../../theme/themed-styles' -import { useMobileColorScheme } from '../../theme/use-mobile-color-scheme' -import type { MobileNewTabAgentOption } from '../agent-launch/mobile-new-tab-agent-options' - -const AGENT_GRID_COLUMNS = 3 - -type Props = { - visible: boolean - workspaceName: string - agentOptions: readonly MobileNewTabAgentOption[] - /** Loading, none, other-server or error; shown instead of the agent grid. */ - agentStatusRow: ActionSheetAction | null - canCreateBrowser: boolean - canCreateMarkdownNote: boolean - onClose: () => void - onCreateAgent: (agent: TuiAgent) => void - onCreateTerminal: () => void - /** Runs after the sheet has gone, so it may open another modal. */ - onCreateBrowser: () => void - onCreateMarkdownNote: () => void -} - -export function MobileNewTabSheet({ - visible, - workspaceName, - agentOptions, - agentStatusRow, - canCreateBrowser, - canCreateMarkdownNote, - onClose, - onCreateAgent, - onCreateTerminal, - onCreateBrowser, - onCreateMarkdownNote -}: Props) { - useMobileColorScheme() - const t = useMobileTranslation() - // Wait for the iOS sheet window to unmount before presenting another native modal. - const afterCloseRef = useRef<(() => void) | null>(null) - const closeThen = (action: () => void) => { - afterCloseRef.current = action - onClose() - } - const closeAndRun = (action: () => void) => { - onClose() - action() - } - const showAgentGrid = agentStatusRow === null && agentOptions.length > 0 - // Short last rows keep empty slots so every tile stays a third of the row wide. - const agentRows: { options: MobileNewTabAgentOption[]; emptySlots: string[] }[] = [] - for (let start = 0; start < agentOptions.length; start += AGENT_GRID_COLUMNS) { - const options = agentOptions.slice(start, start + AGENT_GRID_COLUMNS) - const emptySlots: string[] = [] - for (let column = options.length; column < AGENT_GRID_COLUMNS; column += 1) { - emptySlots.push(`empty:${column}`) - } - agentRows.push({ options, emptySlots }) - } - const iconColor = colors.textSecondary - return ( - { - const run = afterCloseRef.current - afterCloseRef.current = null - run?.() - }} - dragContentToDismiss - > - - {showAgentGrid || agentStatusRow ? ( - - ) : null} - {agentStatusRow ? : null} - {showAgentGrid ? ( - - {agentRows.map((row) => ( - option.agent).join(':')} style={styles.gridRow}> - {row.options.map((option) => ( - closeAndRun(() => onCreateAgent(option.agent))} - accessibilityRole="button" - accessibilityLabel={option.label} - style={styles.tile} - > - {({ pressed }) => ( - <> - {pressed ? : null} - - - {option.label} - - - )} - - ))} - {row.emptySlots.map((slot) => ( - - ))} - - ))} - - ) : null} - - } - onPress={() => closeAndRun(onCreateTerminal)} - /> - {canCreateBrowser ? ( - } - navigates - onPress={() => closeThen(onCreateBrowser)} - /> - ) : null} - {canCreateMarkdownNote ? ( - } - onPress={() => closeAndRun(onCreateMarkdownNote)} - /> - ) : null} - - - ) -} - -const styles = themedStyleSheet(() => ({ - grid: { gap: spacing.sm }, - gridRow: { flexDirection: 'row', gap: spacing.sm }, - tile: { - flex: 1, - minWidth: 0, - height: 78, - borderRadius: radii.surface, - backgroundColor: colors.sheetGroup, - overflow: 'hidden', - alignItems: 'center', - justifyContent: 'center', - gap: spacing.sm, - paddingHorizontal: spacing.sm - }, - // Same wash as a pressed ListRow, laid over the tile so its own fill stays underneath. - tilePressed: { - position: 'absolute', - top: 0, - right: 0, - bottom: 0, - left: 0, - backgroundColor: colors.translucentFill - }, - tileLabel: { fontSize: 13.5, fontWeight: '600', color: colors.textPrimary, maxWidth: '100%' }, - // Padded like a tile: flex shares space after padding, so a bare spacer would widen its row's tiles. - tileSpacer: { flex: 1, paddingHorizontal: spacing.sm }, - actions: { marginTop: spacing.lg } -})) diff --git a/mobile/src/session/tabs/MobileSessionTabBar.tsx b/mobile/src/session/tabs/MobileSessionTabBar.tsx deleted file mode 100644 index a9aeb494c8e..00000000000 --- a/mobile/src/session/tabs/MobileSessionTabBar.tsx +++ /dev/null @@ -1,361 +0,0 @@ -import { useMemo, useRef, useState } from 'react' -import { Pressable, Text, View } from 'react-native' -import { Plus } from 'lucide-react-native' -import { Gesture, GestureDetector } from 'react-native-gesture-handler' -import Animated, { - Easing, - ReduceMotion, - useAnimatedStyle, - useSharedValue, - withSequence, - withTiming -} from 'react-native-reanimated' -import { triggerMediumImpact, triggerSelection } from '../../platform/haptics' -import { useMobileTranslation } from '../../i18n/use-mobile-translation' -import { AgentStateDot } from '../../components/AgentStateDot' -import { ArrivalFade } from '../../components/ArrivalFade' -import { useDrawerReducedMotion } from '../../components/bottom-drawer/bottom-drawer-constants' -import { SHEET_CURVE } from '../../components/sheet-motion-values' -import { useReducedMotionEnabled } from '../../platform/use-reduced-motion' -import { useKeyboardPersistingTaps } from '../../platform/keyboard-persisting-taps' -import { useNow } from '../../hooks/use-now' -import { agentStateLabel, type AgentDotState } from '../../worktree/agent-row-display' -import { colors, feedback, radii, spacing } from '../../theme/mobile-theme' -import { themedStyleSheet } from '../../theme/themed-styles' -import { useMobileColorScheme } from '../../theme/use-mobile-color-scheme' -import { buildMobileSessionTabRow, type MobileSessionTabRow } from './mobile-session-tab-rows' -import { MobileSessionTabMark } from './MobileSessionTabMark' -import { MobileSessionTabsSheet } from './MobileSessionTabsSheet' -import type { MobileSessionController } from '../use-mobile-session-controller' - -const SWIPE_SETTLE_MS = 250 -// A tab arriving from the side starts this far out, as a share of the tab cell's width. -const SWIPE_ENTRY_SHARE = 0.4 -const FLICK_VELOCITY = 600 -// With no tab on that side, the tab gives only this share of the drag. -const EDGE_RESISTANCE = 0.25 -const DRAG_CLICK_WINDOW_MS = 400 - -// A tab drawn before it had any agent state (from the phone's store) fades its dot in when one lands. -function TabBarDot({ tabId, state }: { tabId: string; state: AgentDotState | undefined }) { - const [startedEmpty] = useState({ tabId, empty: state === undefined }) - if (startedEmpty.tabId === tabId && startedEmpty.empty) { - return - } - return state && state !== 'idle' ? : null -} - -// Mounted with the bare tab, so the motion preference is read before the dot arrives. -function ArrivingTabBarDot({ state }: { state: AgentDotState | undefined }) { - const reducedMotion = useReducedMotionEnabled() - if (!state || state === 'idle') { - return null - } - return ( - - - - ) -} - -/** The tab you're in as one cell: tap for its actions, swipe for the next one; the count shows every tab, + adds one. */ -export function MobileSessionTabBar({ - controller, - inTopBar, - actionsCollapsed = false -}: { - controller: MobileSessionController - inTopBar: boolean - actionsCollapsed?: boolean -}) { - useMobileColorScheme() - const t = useMobileTranslation() - const { - activeSessionTabId, - connState, - createTabBusy, - handleOpenNewTab, - switchSessionTab, - openSessionTabActionSheetAfterKeyboardDismiss, - visibleTabs, - terminals - } = controller - // Why: the first tap must reach the bar with the keyboard open instead of only dismissing it (#5106). - const keepsKeyboard = useKeyboardPersistingTaps('handled') - const now = useNow(60_000) - const rows = useMemo( - () => visibleTabs.map((tab) => buildMobileSessionTabRow(tab, terminals, now)), - [visibleTabs, terminals, now] - ) - const activeIndex = rows.findIndex((row) => row.tab.id === activeSessionTabId) - const active: MobileSessionTabRow | undefined = rows[activeIndex] - const previous = activeIndex > 0 ? rows[activeIndex - 1] : undefined - const next = activeIndex !== -1 ? rows[activeIndex + 1] : undefined - const othersNeedYou = rows.some( - (row) => row.tab.id !== activeSessionTabId && row.group === 'needs-you' - ) - const [sheetOpen, setSheetOpen] = useState(false) - - const reducedMotion = useDrawerReducedMotion() - const offset = useSharedValue(0) - const width = useSharedValue(0) - // On the web, releasing a mouse drag also clicks the tab. - const dragEndedAtRef = useRef(0) - const settle = (to: number) => - withTiming( - to, - reducedMotion - ? { duration: 0, reduceMotion: ReduceMotion.Never } - : { duration: SWIPE_SETTLE_MS, easing: Easing.bezier(...SHEET_CURVE) } - ) - const step = (target: MobileSessionTabRow | undefined, direction: 1 | -1) => { - if (!target) { - offset.value = settle(0) - return - } - triggerSelection() - switchSessionTab(target.tab) - offset.value = withSequence( - withTiming(-direction * width.value * SWIPE_ENTRY_SHARE, { duration: 0 }), - settle(0) - ) - } - // Callbacks run on JS: the session page is a web document, and stepping switches React state. - const pan = Gesture.Pan() - .runOnJS(true) - .enabled(rows.length > 1) - .activeOffsetX([-12, 12]) - .failOffsetY([-10, 10]) - .onUpdate((event) => { - const target = event.translationX > 0 ? previous : next - offset.value = target ? event.translationX : event.translationX * EDGE_RESISTANCE - }) - .onEnd((event) => { - // oxlint-disable-next-line react/purity -- runs when the drag ends, never during render. - dragEndedAtRef.current = Date.now() - const direction = event.translationX > 0 ? 1 : -1 - const flicked = - Math.abs(event.velocityX) > FLICK_VELOCITY && Math.sign(event.velocityX) === direction - if (Math.abs(event.translationX) > width.value / 2 || flicked) { - step(direction === 1 ? previous : next, direction) - return - } - offset.value = settle(0) - }) - // Web has no worklet plugin; mapper hooks need this explicit dependency list. - const contentStyle = useAnimatedStyle( - () => ({ - transform: [{ translateX: offset.value }], - opacity: 1 - Math.min(Math.abs(offset.value) / Math.max(width.value, 1), 1) * 0.6 - }), - [offset, width] - ) - - const status = active?.dotState && active.dotState !== 'idle' ? active.dotState : undefined - const countLabel = - rows.length === 1 - ? t('session.tabBar.oneTab', '1 tab') - : t('session.tabBar.tabCount', '{{count}} tabs', { count: rows.length }) - const tabLabel = [active?.title, status ? agentStateLabel(status) : null] - .filter(Boolean) - .join(', ') - const countAccessibilityLabel = [ - countLabel, - othersNeedYou ? t('session.tabBar.anotherNeedsYou', 'another tab needs you') : null - ] - .filter(Boolean) - .join(', ') - const createDisabled = createTabBusy || connState !== 'connected' - const openTabActions = () => { - // The click a drag release produces on the web belongs to the drag. - if (!active || Date.now() - dragEndedAtRef.current < DRAG_CLICK_WINDOW_MS) { - return - } - openSessionTabActionSheetAfterKeyboardDismiss(active.tab) - } - - return ( - - {/* pan-y keeps vertical touches with the page on the web. */} - - { - width.value = event.nativeEvent.layout.width - }} - > - [styles.tabPress, pressed && styles.cellPressed]} - onPress={openTabActions} - onLongPress={() => { - triggerMediumImpact() - openTabActions() - }} - delayLongPress={400} - disabled={!active} - accessibilityRole="button" - accessibilityLabel={tabLabel || countLabel} - accessibilityHint={t('session.tabBar.actionsHint', 'Shows actions for this tab')} - accessibilityActions={[ - ...(next ? [{ name: 'nextTab', label: t('session.tabBar.next', 'Next tab') }] : []), - ...(previous - ? [{ name: 'previousTab', label: t('session.tabBar.previous', 'Previous tab') }] - : []) - ]} - onAccessibilityAction={(event) => { - const name = event.nativeEvent.actionName - if (name === 'nextTab') { - step(next, -1) - } else if (name === 'previousTab') { - step(previous, 1) - } - }} - > - {/* Only the contents slide, so the cell stays flush with the screen edge. */} - - {active ? ( - <> - - - - - - ) : null} - - {active?.title ?? countLabel} - - - - - - [ - styles.sideCell, - actionsCollapsed && !inTopBar && styles.lastCell, - pressed && styles.cellPressed - ]} - onPress={() => setSheetOpen(true)} - accessibilityRole="button" - accessibilityLabel={countAccessibilityLabel} - accessibilityHint={t('session.tabBar.hint', 'Shows all tabs')} - > - - {rows.length} - - {othersNeedYou ? : null} - - {/* A collapsed bar's + lives in the header's ⋯ menu. */} - {!actionsCollapsed ? ( - [ - styles.sideCell, - !inTopBar && styles.lastCell, - pressed && styles.cellPressed, - createDisabled && styles.cellDisabled - ]} - disabled={createDisabled} - onPress={handleOpenNewTab} - accessibilityRole="button" - accessibilityLabel={t('session.tabStrip.newTab', 'New tab')} - accessibilityState={{ disabled: createDisabled }} - > - - - ) : null} - setSheetOpen(false)} - /> - - ) -} - -const styles = themedStyleSheet(() => ({ - bar: { - height: 44, - flexDirection: 'row', - alignItems: 'stretch', - borderTopWidth: 1, - borderTopColor: colors.borderSubtle - }, - // Cancels the top bar's vertical padding so the cells run the bar's full height. - barInTopBar: { - flex: 1, - minWidth: 0, - flexDirection: 'row', - alignItems: 'stretch', - alignSelf: 'stretch', - marginVertical: -spacing.xs, - marginLeft: spacing.md, - borderLeftWidth: 1, - borderLeftColor: colors.borderSubtle - }, - // The open tab, drawn like the strip's active tab: raised, with a line under it. - tabCell: { - flex: 1, - minWidth: 0, - backgroundColor: colors.bgRaised, - borderBottomWidth: 2, - borderBottomColor: colors.textSecondary, - borderRightWidth: 1, - borderRightColor: colors.borderSubtle, - overflow: 'hidden' - }, - tabPress: { - flex: 1, - justifyContent: 'center', - paddingLeft: spacing.lg, - paddingRight: spacing.sm - }, - tabContent: { flexDirection: 'row', alignItems: 'center', gap: 6 }, - // Keeps web SVGs from shrinking before the title ellipsizes. - mark: { flexShrink: 0 }, - title: { - flexShrink: 1, - minWidth: 0, - fontSize: 14, - fontWeight: '600', - color: colors.textPrimary - }, - sideCell: { - width: 46, - alignItems: 'center', - justifyContent: 'center', - borderRightWidth: 1, - borderRightColor: colors.borderSubtle - }, - // At the screen edge; in the top bar the hairline stays to part the cells from the round buttons. - lastCell: { borderRightWidth: 0 }, - cellPressed: { backgroundColor: colors.translucentFill }, - cellDisabled: { opacity: feedback.disabledOpacity }, - countBox: { - minWidth: 19, - height: 19, - paddingHorizontal: 3, - borderRadius: 6, - borderWidth: 1.8, - borderColor: colors.textPrimary, - alignItems: 'center', - justifyContent: 'center' - }, - countText: { - fontSize: 11, - fontWeight: '700', - color: colors.textPrimary, - fontVariant: ['tabular-nums'] - }, - // Another tab needs you: a red dot on the count's corner, ringed in the bar's colour. - alert: { - position: 'absolute', - top: 8, - right: 9, - width: 9, - height: 9, - borderRadius: radii.pill, - backgroundColor: colors.statusRed, - borderWidth: 2, - borderColor: colors.bgPanel - } -})) diff --git a/mobile/src/session/tabs/MobileSessionTabMark.tsx b/mobile/src/session/tabs/MobileSessionTabMark.tsx deleted file mode 100644 index 67a1d0b3895..00000000000 --- a/mobile/src/session/tabs/MobileSessionTabMark.tsx +++ /dev/null @@ -1,28 +0,0 @@ -import { File, FileText, Globe, SquareTerminal } from 'lucide-react-native' -import { MobileAgentIcon } from '../../components/MobileAgentIcon' -import { colors } from '../../theme/mobile-theme' -import { useMobileColorScheme } from '../../theme/use-mobile-color-scheme' -import type { MobileSessionTabRow } from './mobile-session-tab-rows' - -/** The tab's agent icon, else an icon for what kind of tab it is. */ -export function MobileSessionTabMark({ - row, - size -}: { - row: Pick - size: number -}) { - useMobileColorScheme() - if (row.agentId) { - return - } - const Icon = - row.tab.type === 'browser' - ? Globe - : row.tab.type === 'markdown' - ? FileText - : row.tab.type === 'file' - ? File - : SquareTerminal - return -} diff --git a/mobile/src/session/tabs/MobileSessionTabStrip.tsx b/mobile/src/session/tabs/MobileSessionTabStrip.tsx new file mode 100644 index 00000000000..6455c9bf6f8 --- /dev/null +++ b/mobile/src/session/tabs/MobileSessionTabStrip.tsx @@ -0,0 +1,327 @@ +import { useCallback, useMemo, useRef, useState } from 'react' +import { View, Text, ScrollView, Pressable, StyleSheet } from 'react-native' +import { File, FileText, Globe, Plus, SquareTerminal } from 'lucide-react-native' +import { triggerMediumImpact } from '../../platform/haptics' +import { useMobileTranslation } from '../../i18n/use-mobile-translation' +import { MobileAgentIcon } from '../../components/MobileAgentIcon' +import { AgentStateDot } from '../../components/AgentStateDot' +import { ArrivalFade } from '../../components/ArrivalFade' +import { useReducedMotionEnabled } from '../../platform/use-reduced-motion' +import { useNow } from '../../hooks/use-now' +import { + agentDotState, + agentStateLabel, + type AgentDotState +} from '../../worktree/agent-row-display' +import { + getMobileTerminalTabHostAgentIdentity, + getMobileSessionTabTitle, + resolveMobileTerminalTabAgentId +} from '../../terminal/session/mobile-terminal-tab-agent' +import { colors, spacing } from '../../theme/mobile-theme' +import { useMobileColorScheme } from '../../theme/use-mobile-color-scheme' +import { NavBarButton } from '../../components/nav-bar/NavBarButton' +import { + resolveTabStripEdgeHint, + resolveTabStripHiddenTabs, + resolveTabStripTapOffset, + type TabStripHiddenTabs +} from './tab-strip-scroll' +import { styles } from '../mobile-session-styles' +import { useKeyboardPersistingTaps } from '../../platform/keyboard-persisting-taps' +import type { MobileSessionController } from '../use-mobile-session-controller' + +const DOUBLE_TAP_RENAME_MS = 300 + +// A tab drawn before it had any agent state (a strip from the phone's store) fades its dot in when one lands. +function SessionTabAgentDot({ state }: { state: AgentDotState | undefined }) { + const [startedEmpty] = useState(state === undefined) + if (startedEmpty) { + return + } + return state && state !== 'idle' ? : null +} + +// Mounted with the bare tab, so the motion preference is read before the dot arrives. +function ArrivingSessionTabAgentDot({ state }: { state: AgentDotState | undefined }) { + const reducedMotion = useReducedMotionEnabled() + if (!state || state === 'idle') { + return null + } + return ( + + + + ) +} + +export function MobileSessionTabStrip({ + controller, + inTopBar, + actionsCollapsed = false +}: { + controller: MobileSessionController + inTopBar: boolean + actionsCollapsed?: boolean +}) { + useMobileColorScheme() + const t = useMobileTranslation() + const { + activeSessionTabId, + activeSessionTabIdRef, + tabStripRef, + tabStripOffsetRef, + tabStripViewportWidthRef, + tabStripContentWidthRef, + tabLayoutsRef, + connState, + createTabBusy, + handleOpenNewTab, + scrollActiveTabIntoView, + holdTabStripReveal, + cancelHeldTabStripReveal, + switchSessionTab, + openSessionTabActionSheetAfterKeyboardDismiss, + setRenameTarget, + activeHandleRef, + visibleTabs, + terminals + } = controller + const tabBarKeepsKeyboard = useKeyboardPersistingTaps('handled') + const lastTabTapRef = useRef<{ id: string; at: number } | null>(null) + const now = useNow(60_000) + const tabStates = useMemo(() => { + const states = new Map() + for (const tab of visibleTabs) { + // Why: a status restored across a host restart is unconfirmed, so like desktop it shows no live state. + if (tab.type === 'terminal' && tab.agentStatus && !tab.agentStatus.restoredUnconfirmed) { + states.set(tab.id, agentDotState(tab.agentStatus, now)) + } else if (tab.type === 'agent-session' && tab.chatStatus) { + states.set(tab.id, agentDotState(tab.chatStatus, now)) + } + } + return states + }, [visibleTabs, now]) + const [hiddenTabs, setHiddenTabs] = useState({ left: [], right: [] }) + const refreshHiddenTabs = () => { + const next = resolveTabStripHiddenTabs({ + layouts: tabLayoutsRef.current, + viewportWidth: tabStripViewportWidthRef.current, + currentOffset: tabStripOffsetRef.current + }) + // Why: scroll fires every frame; re-render only when a tab crosses an edge. + setHiddenTabs((prev) => + prev.left.join('\n') === next.left.join('\n') && + prev.right.join('\n') === next.right.join('\n') + ? prev + : next + ) + } + const leftHint = resolveTabStripEdgeHint(hiddenTabs.left, tabStates) + const rightHint = resolveTabStripEdgeHint(hiddenTabs.right, tabStates) + const pressTab = useCallback( + (t: (typeof visibleTabs)[number]) => { + const at = Date.now() + const lastTap = lastTabTapRef.current + lastTabTapRef.current = { id: t.id, at } + if ( + lastTap?.id === t.id && + at - lastTap.at < DOUBLE_TAP_RENAME_MS && + t.type === 'terminal' && + typeof t.terminal === 'string' + ) { + lastTabTapRef.current = null + triggerMediumImpact() + setRenameTarget({ + handle: t.terminal, + title: t.title, + isActive: t.terminal === activeHandleRef.current + }) + // The first tap's held reveal still runs once the window closes. + return + } + cancelHeldTabStripReveal() + const nextOffset = resolveTabStripTapOffset({ + tabId: t.id, + tabs: visibleTabs, + layouts: tabLayoutsRef.current, + viewportWidth: tabStripViewportWidthRef.current, + contentWidth: tabStripContentWidthRef.current, + currentOffset: tabStripOffsetRef.current + }) + if (nextOffset !== tabStripOffsetRef.current) { + holdTabStripReveal(nextOffset, DOUBLE_TAP_RENAME_MS) + } + switchSessionTab(t) + }, + [ + activeHandleRef, + cancelHeldTabStripReveal, + holdTabStripReveal, + setRenameTarget, + switchSessionTab, + tabLayoutsRef, + tabStripContentWidthRef, + tabStripOffsetRef, + tabStripRef, + tabStripViewportWidthRef, + visibleTabs + ] + ) + const attachTabStrip = useCallback( + (view: ScrollView | null) => { + if (view) { + tabStripOffsetRef.current = 0 + tabStripViewportWidthRef.current = 0 + tabStripContentWidthRef.current = 0 + tabLayoutsRef.current.clear() + } + tabStripRef.current = view + }, + [ + tabStripRef, + tabStripOffsetRef, + tabStripViewportWidthRef, + tabStripContentWidthRef, + tabLayoutsRef + ] + ) + return ( + + {/* Why: tab taps must register on first press with the keyboard open instead of being eaten by dismissal (#5106). */} + + { + tabStripOffsetRef.current = e.nativeEvent.contentOffset.x + refreshHiddenTabs() + }} + onLayout={(e) => { + tabStripViewportWidthRef.current = e.nativeEvent.layout.width + scrollActiveTabIntoView(activeSessionTabIdRef.current, false) + refreshHiddenTabs() + }} + onContentSizeChange={(width) => { + tabStripContentWidthRef.current = width + scrollActiveTabIntoView(activeSessionTabIdRef.current, false) + refreshHiddenTabs() + }} + > + {visibleTabs.map((tab, index) => { + const tabState = tabStates.get(tab.id) + const hostAgentIdentity = getMobileTerminalTabHostAgentIdentity(tab, terminals) + const title = getMobileSessionTabTitle(tab, hostAgentIdentity) + return ( + [ + styles.tab, + index === 0 && !inTopBar && styles.tabFirst, + tab.id === activeSessionTabId && styles.tabActive, + pressed && styles.tabPressed + ]} + onLayout={(e) => { + const { x, width } = e.nativeEvent.layout + tabLayoutsRef.current.set(tab.id, { x, width }) + if (tab.id === activeSessionTabIdRef.current) { + scrollActiveTabIntoView(tab.id, false) + } + refreshHiddenTabs() + }} + onPress={() => pressTab(tab)} + accessibilityRole="tab" + accessibilityLabel={ + tabState && tabState !== 'idle' + ? t('session.tabStrip.tabStatus', '{{title}}, {{status}}', { + title, + status: agentStateLabel(tabState) + }) + : title + } + accessibilityState={{ selected: tab.id === activeSessionTabId }} + onLongPress={() => { + triggerMediumImpact() + openSessionTabActionSheetAfterKeyboardDismiss(tab) + }} + delayLongPress={400} + > + + + + {tab.type === 'browser' && ( + + )} + {tab.type === 'markdown' && ( + + )} + {tab.type === 'file' && ( + + )} + {tab.type === 'agent-session' && ( + + )} + {tab.type === 'terminal' && + (() => { + const agentId = resolveMobileTerminalTabAgentId(tab, hostAgentIdentity) + return agentId ? ( + + ) : ( + + ) + })()} + + + {title} + + + + ) + })} + + {leftHint ? ( + + + + ) : null} + {rightHint ? ( + + + + ) : null} + + {/* Terminal layout and Quick commands live in the header's ⋯ menu. */} + {!actionsCollapsed ? ( + + + + ) : null} + + ) +} + +// The right inset matches the nav bar's, so the + lines up with the ⋯ above it. +const newTabStyles = StyleSheet.create({ + slot: { + paddingLeft: spacing.sm, + paddingRight: spacing.md + } +}) diff --git a/mobile/src/session/tabs/MobileSessionTabsSheet.tsx b/mobile/src/session/tabs/MobileSessionTabsSheet.tsx deleted file mode 100644 index f8cd5cd1981..00000000000 --- a/mobile/src/session/tabs/MobileSessionTabsSheet.tsx +++ /dev/null @@ -1,220 +0,0 @@ -import { Fragment, useRef, useState } from 'react' -import { Pressable, Text, View } from 'react-native' -import { Plus, X } from 'lucide-react-native' -import { BottomDrawer } from '../../components/bottom-drawer/BottomDrawer' -import { ListGroup, SectionHeader } from '../../components/list/ListGroup' -import { NavBarButton } from '../../components/nav-bar/NavBarButton' -import { SheetHeader } from '../../components/nav-bar/SheetHeader' -import { SwipeRowActions } from '../../components/SwipeRowActions' -import { AgentStateDot } from '../../components/AgentStateDot' -import { useMobileTranslation } from '../../i18n/use-mobile-translation' -import { useNow } from '../../hooks/use-now' -import { formatTimeAgo } from '../../worktree/agent-row-display' -import { colors, radii, spacing } from '../../theme/mobile-theme' -import { themedStyleSheet } from '../../theme/themed-styles' -import { useMobileColorScheme } from '../../theme/use-mobile-color-scheme' -import type { MobileSessionTabGroup, MobileSessionTabRow } from './mobile-session-tab-rows' -import { MobileSessionTabMark } from './MobileSessionTabMark' -import type { MobileSessionController } from '../use-mobile-session-controller' - -const TILE_SIZE = 34 -const TILE_DOT_RING = 2 -const GROUP_ORDER: readonly MobileSessionTabGroup[] = ['needs-you', 'working', 'other'] - -/** Every tab in the workspace, what needs you first; tap one to go there, swipe one to close it. */ -export function MobileSessionTabsSheet({ - controller, - rows, - visible, - onClose -}: { - controller: MobileSessionController - rows: readonly MobileSessionTabRow[] - visible: boolean - onClose: () => void -}) { - useMobileColorScheme() - const t = useMobileTranslation() - const now = useNow(60_000) - const { - activeSessionTabId, - worktreeName, - switchSessionTab, - handleCloseSessionTab, - handleOpenNewTab, - createTabBusy, - connState - } = controller - // Sections hold still while the sheet is open, so a row never moves under a finger. - const [frozenGroups, setFrozenGroups] = useState>() - if (visible && frozenGroups === undefined) { - setFrozenGroups(new Map(rows.map((row) => [row.tab.id, row.group]))) - } - const afterCloseRef = useRef<(() => void) | null>(null) - const sections = GROUP_ORDER.map((group) => ({ - group, - rows: rows.filter((row) => (frozenGroups?.get(row.tab.id) ?? row.group) === group) - })).filter((section) => section.rows.length > 0) - const sectionTitles: Record = { - 'needs-you': t('session.tabsSheet.needsYou', 'Needs you'), - working: t('session.tabsSheet.working', 'Working'), - other: t('session.tabsSheet.everythingElse', 'Everything else') - } - const title = - rows.length === 1 - ? t('session.tabBar.oneTab', '1 tab') - : t('session.tabBar.tabCount', '{{count}} tabs', { count: rows.length }) - - const renderRow = (row: MobileSessionTabRow) => { - const current = row.tab.id === activeSessionTabId - const badge = row.dotState && row.dotState !== 'idle' ? row.dotState : undefined - const time = current - ? t('session.tabsSheet.open', 'Open') - : row.timeAt !== null - ? formatTimeAgo(row.timeAt, now) - : '' - return ( - void handleCloseSessionTab(row.tab) - }} - > - {(guardPress) => ( - { - if (!current) { - switchSessionTab(row.tab) - } - onClose() - })} - > - {({ pressed }) => ( - <> - {current || pressed ? : null} - - - {badge ? ( - - - - ) : null} - - - - {row.title} - - - {row.detail} - - - {time ? ( - {time} - ) : null} - - )} - - )} - - ) - } - - return ( - { - setFrozenGroups(undefined) - const afterClose = afterCloseRef.current - afterCloseRef.current = null - afterClose?.() - }} - > - - - - { - // The new-tab sheet is its own drawer; open it once this one is gone. - afterCloseRef.current = handleOpenNewTab - onClose() - }} - /> - - - - {sections.map((section, index) => ( - - {sections.length > 1 ? ( - - ) : null} - - {section.rows.map(renderRow)} - - - ))} - - - ) -} - -const styles = themedStyleSheet(() => ({ - // Over the sheet header's empty trailing side, centred in its 52pt row. - newTab: { position: 'absolute', top: 8, right: spacing.xs }, - body: { paddingBottom: spacing.lg }, - // Opaque so a swipe's action stays hidden until the row slides off it. - row: { - minHeight: 60, - flexDirection: 'row', - alignItems: 'center', - gap: spacing.md, - paddingHorizontal: spacing.lg, - paddingVertical: spacing.sm, - backgroundColor: colors.sheetGroup - }, - // The tab you're in, or a pressed row, washed like a picked segment on a sheet group. - wash: { - position: 'absolute', - top: 0, - right: 0, - bottom: 0, - left: 0, - backgroundColor: colors.translucentFill - }, - tile: { - width: TILE_SIZE, - height: TILE_SIZE, - borderRadius: radii.tile, - backgroundColor: colors.translucentFill, - alignItems: 'center', - justifyContent: 'center' - }, - // The state dot on the tile's corner, ringed in the group's colour. - tileBadge: { - position: 'absolute', - top: -4, - right: -4, - padding: TILE_DOT_RING, - borderRadius: radii.pill, - backgroundColor: colors.sheetGroup - }, - text: { flex: 1, minWidth: 0, gap: 2 }, - title: { fontSize: 15, fontWeight: '600', color: colors.textPrimary }, - detail: { fontSize: 13, color: colors.textSecondary }, - time: { fontSize: 12, color: colors.textMuted, fontVariant: ['tabular-nums'] }, - timeCurrent: { fontWeight: '600', color: colors.accentBlue } -})) diff --git a/mobile/src/session/tabs/mobile-session-tab-rows.ts b/mobile/src/session/tabs/mobile-session-tab-rows.ts deleted file mode 100644 index 1b2854512f6..00000000000 --- a/mobile/src/session/tabs/mobile-session-tab-rows.ts +++ /dev/null @@ -1,121 +0,0 @@ -import { agentListPreview } from '../../agents/agent-list-row-text' -import { agentStatusGroup } from '../../agents/agent-list-rows' -import { translate } from '../../i18n/mobile-i18n' -import { - getMobileSessionTabTitle, - getMobileTerminalTabHostAgentIdentity, - resolveMobileTerminalTabAgentId -} from '../../terminal/session/mobile-terminal-tab-agent' -import { - agentDotState, - agentRowTimeAt, - agentStateLabel, - type AgentDotState -} from '../../worktree/agent-row-display' -import type { MobileSessionTab } from './mobile-session-tabs' - -export type MobileSessionTabGroup = 'needs-you' | 'working' | 'other' - -export type MobileSessionTabRow = { - tab: MobileSessionTab - title: string - /** The agent whose icon marks the tab; null draws the tab kind's icon. */ - agentId: string | null - dotState: AgentDotState | undefined - group: MobileSessionTabGroup - /** The second line: the state, then what the tab is doing or showing. */ - detail: string - timeAt: number | null -} - -export function mobileSessionTabDotState( - tab: MobileSessionTab, - now: number -): AgentDotState | undefined { - // Why: a status restored across a host restart is unconfirmed, so like desktop it shows no live state. - if (tab.type === 'terminal' && tab.agentStatus && !tab.agentStatus.restoredUnconfirmed) { - return agentDotState(tab.agentStatus, now) - } - if (tab.type === 'agent-session' && tab.chatStatus) { - return agentDotState(tab.chatStatus, now) - } - return undefined -} - -function tabGroup(state: AgentDotState | undefined): MobileSessionTabGroup { - if (!state) { - return 'other' - } - const group = agentStatusGroup(state) - return group === 'needs-you' || group === 'working' ? group : 'other' -} - -function tabContent( - tab: MobileSessionTab, - title: string, - dotState: AgentDotState | undefined -): string { - switch (tab.type) { - case 'terminal': { - const status = tab.agentStatus - if (!status || !dotState) { - return '' - } - return agentListPreview(status, agentStatusGroup(dotState), title) - } - case 'browser': - return tab.url === title ? '' : tab.url - case 'markdown': - case 'file': - return tab.relativePath === title ? '' : tab.relativePath - case 'agent-session': - return '' - } -} - -function tabKindLabel(tab: MobileSessionTab): string { - switch (tab.type) { - case 'browser': - return translate('session.tabRows.browser', 'Browser') - case 'markdown': - return translate('session.tabRows.markdown', 'Markdown note') - case 'file': - return translate('session.tabRows.file', 'File') - case 'agent-session': - return translate('session.tabRows.chat', 'Chat') - case 'terminal': - return translate('session.tabRows.terminal', 'Terminal') - } -} - -export function buildMobileSessionTabRow( - tab: MobileSessionTab, - terminals: readonly { handle: string; agentIdentity?: string }[], - now: number -): MobileSessionTabRow { - const hostAgentIdentity = getMobileTerminalTabHostAgentIdentity(tab, terminals) - const title = getMobileSessionTabTitle(tab, hostAgentIdentity) - const dotState = mobileSessionTabDotState(tab, now) - const stateWord = dotState && dotState !== 'idle' ? agentStateLabel(dotState) : '' - const content = tabContent(tab, title, dotState) - const timeAt = - tab.type === 'terminal' && tab.agentStatus && !tab.agentStatus.restoredUnconfirmed - ? agentRowTimeAt(tab.agentStatus) - : tab.type === 'agent-session' && tab.chatStatus - ? tab.chatStatus.updatedAt - : null - return { - tab, - title, - agentId: - tab.type === 'agent-session' - ? tab.agent - : tab.type === 'terminal' - ? resolveMobileTerminalTabAgentId(tab, hostAgentIdentity) - : null, - dotState, - group: tabGroup(dotState), - detail: [stateWord, content].filter(Boolean).join(' · ') || tabKindLabel(tab), - timeAt - } -} diff --git a/mobile/src/session/tabs/tab-strip-scroll.ts b/mobile/src/session/tabs/tab-strip-scroll.ts index e7d6c85e416..9b9dd9ce99b 100644 --- a/mobile/src/session/tabs/tab-strip-scroll.ts +++ b/mobile/src/session/tabs/tab-strip-scroll.ts @@ -1,3 +1,5 @@ +import type { AgentDotState } from '../../worktree/agent-row-display' + export type TabStripScrollInput = { tabX: number tabWidth: number @@ -34,3 +36,98 @@ export function resolveTabStripScrollOffset({ return Math.min(Math.max(0, nextOffset), maxOffset) } + +type TabLayouts = ReadonlyMap + +export function resolveTabStripTapOffset(args: { + tabId: string + tabs: readonly { id: string }[] + layouts: TabLayouts + viewportWidth: number + contentWidth: number + currentOffset: number +}): number { + const { tabs, layouts, viewportWidth, contentWidth, currentOffset } = args + const index = tabs.findIndex((tab) => tab.id === args.tabId) + const selected = layouts.get(args.tabId) + if (index === -1 || !selected || viewportWidth <= 0) { + return currentOffset + } + const selectedOnly = resolveTabStripScrollOffset({ + tabX: selected.x, + tabWidth: selected.width, + viewportWidth, + contentWidth, + currentOffset + }) + if (selectedOnly === currentOffset) { + return currentOffset + } + const left = selected.x + selected.width / 2 < currentOffset + viewportWidth / 2 + const neighbor = tabs[index + (left ? -1 : 1)] + const layout = neighbor ? layouts.get(neighbor.id) : undefined + let start = selected.x + let end = selected.x + selected.width + if (layout) { + const gap = left ? selected.x - (layout.x + layout.width) : layout.x - end + // Keep the tapped tab whole while revealing as much of its neighbor as fits within both margins. + const peek = Math.min(layout.width, Math.max(0, viewportWidth - 24 - selected.width - gap)) + if (left) { + start -= gap + peek + } else { + end += gap + peek + } + } + return resolveTabStripScrollOffset({ + tabX: start, + tabWidth: end - start, + viewportWidth, + contentWidth, + currentOffset + }) +} + +export type TabStripHiddenTabs = { left: readonly string[]; right: readonly string[] } + +export function resolveTabStripHiddenTabs(args: { + layouts: TabLayouts + viewportWidth: number + currentOffset: number +}): TabStripHiddenTabs { + const left: string[] = [] + const right: string[] = [] + if (args.viewportWidth > 0) { + for (const [tabId, layout] of args.layouts) { + const center = layout.x + layout.width / 2 + if (center < args.currentOffset) { + left.push(tabId) + } else if (center > args.currentOffset + args.viewportWidth) { + right.push(tabId) + } + } + } + return { left, right } +} + +// Lower ranks win when several hidden tabs share an edge; other states get no edge hint. +const EDGE_HINT_RANK: Partial> = { + blocked: 0, + waiting: 0, + working: 1, + monitoring: 1 +} + +export function resolveTabStripEdgeHint( + tabIds: readonly string[], + tabStates: ReadonlyMap +): AgentDotState | null { + let hint: AgentDotState | null = null + for (const tabId of tabIds) { + const state = tabStates.get(tabId) + const rank = state ? EDGE_HINT_RANK[state] : undefined + if (state && rank !== undefined && (hint === null || rank < (EDGE_HINT_RANK[hint] ?? 0))) { + hint = state + } + } + return hint +} diff --git a/mobile/src/session/use-mobile-session-foundation.ts b/mobile/src/session/use-mobile-session-foundation.ts index cdc21bbabb0..6ecb5ae9870 100644 --- a/mobile/src/session/use-mobile-session-foundation.ts +++ b/mobile/src/session/use-mobile-session-foundation.ts @@ -52,7 +52,6 @@ export function useMobileSessionFoundation() { name: worktreeName, path: worktreePath, hostId: worktreeHostId, - branch: worktreeBranch, resolution: worktreeResolution } = useLiveWorktreeName({ client, @@ -116,7 +115,6 @@ export function useMobileSessionFoundation() { worktreeName, worktreePath, worktreeHostId, - worktreeBranch, worktreeResolution, isWideLayout, activePanel, diff --git a/mobile/src/session/use-mobile-session-panel-route-actions.tsx b/mobile/src/session/use-mobile-session-panel-route-actions.tsx index c4ae529e4a8..6cd74e382eb 100644 --- a/mobile/src/session/use-mobile-session-panel-route-actions.tsx +++ b/mobile/src/session/use-mobile-session-panel-route-actions.tsx @@ -7,7 +7,7 @@ import { shouldShowSessionHeaderChecksAction, panelRouteDescriptor } from './session-panel-host' -import type { ActionSheetAction } from '../components/ActionSheetModal' +import { MobileAgentIcon } from '../components/MobileAgentIcon' import { useMobileTranslation } from '../i18n/use-mobile-translation' import { useWorkspaceExecutionHost, @@ -49,81 +49,127 @@ export function useMobileSessionPanelRouteActions(scope: MobileSessionPresentati quickCommandsSupported, showToast } = scope - const agentStatusRow = (errorHint: string): ActionSheetAction | null => + const otherRuntimeAgentRow = { + id: 'otherServerAgents', + label: t('session.usePanelRouteActions.otherServerAgents', 'Agents on Another Server'), + hint: t('session.usePanelRouteActions.pairServerHint', 'Pair that server directly'), + icon: Bot, + disabled: true, + onPress: () => {} + } + const createTabAgentActions = createTabAgentLoadState === 'loading' - ? { - id: 'detectingAgents', - label: t('session.usePanelRouteActions.detectingAgents', 'Detecting Agents'), - icon: Bot, - disabled: true, - loading: true, - onPress: () => {} - } + ? [ + { + id: 'detectingAgents', + label: t('session.usePanelRouteActions.detectingAgents', 'Detecting Agents'), + icon: Bot, + disabled: true, + loading: true, + onPress: () => {} + } + ] : createTabAgentOptions.length > 0 - ? null - : createTabAgentLoadState === 'loaded' - ? { - id: 'noEnabledAgents', - label: t('session.usePanelRouteActions.noEnabledAgents', 'No Enabled Agents'), - icon: Bot, - disabled: true, - onPress: () => {} + ? createTabAgentOptions.map((option) => ({ + id: `agent:${option.agent}`, + label: option.label, + renderIcon: () => , + onPress: () => { + setShowCreateTabDrawer(false) + void handleCreateTerminal(option.agent) } - : createTabAgentLoadState === 'other-runtime' - ? { - id: 'otherServerAgents', - label: t( - 'session.usePanelRouteActions.otherServerAgents', - 'Agents on Another Server' - ), - hint: t('session.usePanelRouteActions.pairServerHint', 'Pair that server directly'), + })) + : createTabAgentLoadState === 'loaded' + ? [ + { + id: 'noEnabledAgents', + label: t('session.usePanelRouteActions.noEnabledAgents', 'No Enabled Agents'), icon: Bot, disabled: true, onPress: () => {} } + ] + : createTabAgentLoadState === 'other-runtime' + ? [otherRuntimeAgentRow] : createTabAgentLoadState === 'error' - ? { - id: 'presetsUnavailable', - label: t( - 'session.usePanelRouteActions.presetsUnavailable', - 'Agent Presets Unavailable' - ), - hint: errorHint, + ? [ + { + id: 'presetsUnavailable', + label: t( + 'session.usePanelRouteActions.presetsUnavailable', + 'Agent Presets Unavailable' + ), + hint: t( + 'session.usePanelRouteActions.checkHostHint', + 'Check the host connection' + ), + icon: Bot, + disabled: true, + onPress: () => {} + } + ] + : [] + const sendDiffNotesAgentActions = + pendingDiffNotesDelivery === null + ? [] + : createTabAgentLoadState === 'loading' + ? [ + { + id: 'detectingAgents', + label: t('session.usePanelRouteActions.detectingAgents', 'Detecting Agents'), + icon: Bot, + disabled: true, + loading: true, + onPress: () => {} + } + ] + : createTabAgentOptions.length > 0 + ? createTabAgentOptions.map((option) => ({ + id: `agent:${option.agent}`, + label: option.label, + hint: t('session.usePanelRouteActions.newSessionHint', 'New agent session'), + icon: Bot, + onPress: () => { + const delivery = pendingDiffNotesDelivery + setPendingDiffNotesDelivery(null) + if (!delivery) { + return + } + void sendDiffNotesToNewAgent(delivery, async () => { + await handleCreateTerminal(option.agent, { + initialPrompt: delivery.prompt, + onPromptSent: () => void clearDeliveredDiffComments(delivery.comments) + }) + }) + } + })) + : createTabAgentLoadState === 'loaded' + ? [ + { + id: 'noEnabledAgents', + label: t('session.usePanelRouteActions.noEnabledAgents', 'No Enabled Agents'), icon: Bot, disabled: true, onPress: () => {} } - : null - const createTabAgentStatusRow = agentStatusRow( - t('session.usePanelRouteActions.checkHostHint', 'Check the host connection') - ) - const sendDiffNotesStatusRow = agentStatusRow( - t('session.usePanelRouteActions.copyNotesHint', 'Copy notes instead') - ) - const sendDiffNotesAgentActions: ActionSheetAction[] = - pendingDiffNotesDelivery === null - ? [] - : sendDiffNotesStatusRow - ? [sendDiffNotesStatusRow] - : createTabAgentOptions.map((option) => ({ - id: `agent:${option.agent}`, - label: option.label, - hint: t('session.usePanelRouteActions.newSessionHint', 'New agent session'), - icon: Bot, - onPress: () => { - const delivery = pendingDiffNotesDelivery - setPendingDiffNotesDelivery(null) - if (!delivery) { - return - } - void sendDiffNotesToNewAgent(delivery, async () => { - await handleCreateTerminal(option.agent, { - initialPrompt: delivery.prompt, - onPromptSent: () => void clearDeliveredDiffComments(delivery.comments) - }) - }) - } - })) + ] + : createTabAgentLoadState === 'other-runtime' + ? [otherRuntimeAgentRow] + : createTabAgentLoadState === 'error' + ? [ + { + id: 'presetsUnavailable', + label: t( + 'session.usePanelRouteActions.presetsUnavailable', + 'Agent Presets Unavailable' + ), + hint: t('session.usePanelRouteActions.copyNotesHint', 'Copy notes instead'), + icon: Bot, + disabled: true, + onPress: () => {} + } + ] + : [] const handleSessionContentRowLayout = useCallback((event: LayoutChangeEvent) => { const width = Math.round(event.nativeEvent.layout.width) @@ -189,7 +235,7 @@ export function useMobileSessionPanelRouteActions(scope: MobileSessionPresentati ) } return { - createTabAgentStatusRow, + createTabAgentActions, sendDiffNotesAgentActions, handleSessionContentRowLayout, handlePanelTap, diff --git a/mobile/src/session/use-mobile-session-presentation.ts b/mobile/src/session/use-mobile-session-presentation.ts index 8f4b3c23331..374e1db46e8 100644 --- a/mobile/src/session/use-mobile-session-presentation.ts +++ b/mobile/src/session/use-mobile-session-presentation.ts @@ -76,8 +76,11 @@ export function useMobileSessionPresentation(scope: MobileSessionBulkCloseModel) presentedConnState === 'connected' ? showLoadingState && visibleTabs.length === 0 ? t('session.presentation.loadingTabs', 'Loading tabs') - : // The tab count lives in the tab bar, so the header names the branch instead. - (scope.worktreeBranch ?? MOBILE_SESSION_STATUS_LABELS.connected) + : visibleTabs.length === 1 + ? t('session.presentation.oneTab', '1 tab') + : t('session.presentation.tabCount', '{{tabCount}} tabs', { + tabCount: visibleTabs.length + }) : showConnectionRetry ? t('session.presentation.tapToRetry', '{{status}} — tap to retry', { status: verdictDisplayLabel(connectionVerdict) diff --git a/mobile/src/session/use-mobile-session-screen-state.ts b/mobile/src/session/use-mobile-session-screen-state.ts index 45fd4a9d886..bde653c4045 100644 --- a/mobile/src/session/use-mobile-session-screen-state.ts +++ b/mobile/src/session/use-mobile-session-screen-state.ts @@ -1,7 +1,7 @@ import { useState, useRef, useMemo } from 'react' import { useTerminalSplitLayout } from '../terminal/session/use-terminal-split-layout' import { createLiveInputCapture } from '../terminal/session/live-input-capture' -import { Animated } from 'react-native' +import { Animated, type ScrollView } from 'react-native' import type { MobileTerminalLinkOpenMode } from '../storage/preferences' import type { TerminalKeyboardAvoidanceMetrics } from '../terminal/terminal-webview-contract' import { @@ -46,6 +46,11 @@ export function useMobileSessionScreenState(scope: MobileSessionFoundationModel) const activeSessionTabIdRef = useRef(null) // Preserve an explicit phone tab pick while a host snapshot is transiently incomplete. const selectedSessionTabIdRef = useRef(null) + const tabStripRef = useRef(null) + const tabStripOffsetRef = useRef(0) + const tabStripViewportWidthRef = useRef(0) + const tabStripContentWidthRef = useRef(0) + const tabLayoutsRef = useRef>(new Map()) const [markdownDocs, setMarkdownDocs] = useState>(new Map()) const markdownDocsRef = useRef>(new Map()) const [fileDocs, setFileDocs] = useState>(new Map()) @@ -129,6 +134,11 @@ export function useMobileSessionScreenState(scope: MobileSessionFoundationModel) setActiveSessionTabId, activeSessionTabIdRef, selectedSessionTabIdRef, + tabStripRef, + tabStripOffsetRef, + tabStripViewportWidthRef, + tabStripContentWidthRef, + tabLayoutsRef, markdownDocs, setMarkdownDocs, markdownDocsRef, diff --git a/mobile/src/terminal/session/use-mobile-session-keyboard-state.ts b/mobile/src/terminal/session/use-mobile-session-keyboard-state.ts index 785f282bb4f..ec532d01bf5 100644 --- a/mobile/src/terminal/session/use-mobile-session-keyboard-state.ts +++ b/mobile/src/terminal/session/use-mobile-session-keyboard-state.ts @@ -1,9 +1,10 @@ -import { useEffect, useCallback } from 'react' +import { useEffect, useCallback, useRef } from 'react' import { currentSoftKeyboardHeight, subscribeSoftKeyboard } from '../../platform/keyboard-occlusion' import { useTerminalViewportRefit } from '../terminal-viewport-refit' import { removeCustomKey, type CustomKey } from '../terminal-custom-keys' import { useWorkspaceExecutionHost } from '../../navigation/workspace-execution-host' import { writeLastVisitedWorktree } from '../../worktree/last-visited-worktree-by-host' +import { resolveTabStripScrollOffset } from '../../session/tabs/tab-strip-scroll' import type { MobileSessionLifecycleModel } from '../../session/use-mobile-session-lifecycle' export type MobileSessionKeyboardScope = Pick< @@ -11,6 +12,7 @@ export type MobileSessionKeyboardScope = Pick< | 'activeHandleRef' | 'visibleTerminalHandlesRef' | 'terminalPaneViewportsRef' + | 'activeSessionTabId' | 'clientRef' | 'connState' | 'deviceTokenRef' @@ -20,6 +22,11 @@ export type MobileSessionKeyboardScope = Pick< | 'setShowCustomKeyModal' | 'showNativeChatRef' | 'subscribeToTerminal' + | 'tabLayoutsRef' + | 'tabStripContentWidthRef' + | 'tabStripOffsetRef' + | 'tabStripRef' + | 'tabStripViewportWidthRef' | 'terminalFrameRef' | 'terminalRefs' | 'terminals' @@ -38,6 +45,12 @@ export function useMobileSessionKeyboardState(scope: MobileSessionKeyboardScope) connState, terminals, terminalTextScale, + activeSessionTabId, + tabStripRef, + tabStripOffsetRef, + tabStripViewportWidthRef, + tabStripContentWidthRef, + tabLayoutsRef, setCustomKeys, setShowCustomKeyModal, deviceTokenRef, @@ -85,6 +98,55 @@ export function useMobileSessionKeyboardState(scope: MobileSessionKeyboardScope) ) }, [notifyKeyboardVisibility]) + const heldTabStripRevealRef = useRef | null>(null) + const cancelHeldTabStripReveal = useCallback(() => { + if (heldTabStripRevealRef.current) { + clearTimeout(heldTabStripRevealRef.current) + heldTabStripRevealRef.current = null + } + }, []) + useEffect(() => cancelHeldTabStripReveal, [cancelHeldTabStripReveal]) + // Hold strip movement through double-tap detection so its second tap reaches the same tab. + const holdTabStripReveal = useCallback( + (offset: number, delayMs: number) => { + cancelHeldTabStripReveal() + heldTabStripRevealRef.current = setTimeout(() => { + heldTabStripRevealRef.current = null + if (offset !== tabStripOffsetRef.current) { + tabStripOffsetRef.current = offset + tabStripRef.current?.scrollTo({ x: offset, animated: true }) + } + }, delayMs) + }, + [cancelHeldTabStripReveal] + ) + + const scrollActiveTabIntoView = useCallback((tabId: string | null, animated: boolean) => { + if (!tabId || heldTabStripRevealRef.current) { + return + } + const layout = tabLayoutsRef.current.get(tabId) + if (!layout) { + return + } + const nextOffset = resolveTabStripScrollOffset({ + tabX: layout.x, + tabWidth: layout.width, + viewportWidth: tabStripViewportWidthRef.current, + contentWidth: tabStripContentWidthRef.current, + currentOffset: tabStripOffsetRef.current + }) + if (nextOffset !== tabStripOffsetRef.current) { + tabStripOffsetRef.current = nextOffset + tabStripRef.current?.scrollTo({ x: nextOffset, animated }) + } + }, []) + + useEffect(() => { + const id = requestAnimationFrame(() => scrollActiveTabIntoView(activeSessionTabId, true)) + return () => cancelAnimationFrame(id) + }, [activeSessionTabId, scrollActiveTabIntoView]) + const executionHost = useWorkspaceExecutionHost() useEffect(() => { if (hostId && worktreeId) { @@ -109,6 +171,9 @@ export function useMobileSessionKeyboardState(scope: MobileSessionKeyboardScope) notifyTerminalFrameWidth, notifyKeyboardVisibility, notifyTerminalCellBoxChange, + scrollActiveTabIntoView, + holdTabStripReveal, + cancelHeldTabStripReveal, handleDeleteCustomKey, handleManageShortcuts } diff --git a/mobile/src/worktree/agent-row-display.ts b/mobile/src/worktree/agent-row-display.ts index 5fc384ec91d..5969f45afc0 100644 --- a/mobile/src/worktree/agent-row-display.ts +++ b/mobile/src/worktree/agent-row-display.ts @@ -88,7 +88,7 @@ export function agentDisplayLabel(row: RuntimeWorktreeAgentRow, now: number): st // Mirror desktop lastEnteredDoneAt: date main-agent failures independently while subagents remain active. export function agentRowTimeAt( - row: AgentRowVerdictSource & Pick + row: Pick ): number { if (row.state !== 'done' && row.mainAgent && agentVerdictDisplayMark(row) === 'failed') { return row.mainAgent.stateStartedAt diff --git a/mobile/src/worktree/session-worktree-record-reply-schema.ts b/mobile/src/worktree/session-worktree-record-reply-schema.ts index 6ac8c10d1b8..c5b807e093c 100644 --- a/mobile/src/worktree/session-worktree-record-reply-schema.ts +++ b/mobile/src/worktree/session-worktree-record-reply-schema.ts @@ -14,7 +14,6 @@ export const sessionWorktreeRecordSchema = z repo: salvagedOptional('repo', z.string().nullable()), path: salvagedOptional('path', z.string()), hostId: salvagedOptional('hostId', z.string()), - branch: salvagedOptional('branch', z.string().nullable()), diffComments: z.unknown().optional() }) ) @@ -23,11 +22,6 @@ export const sessionWorktreeRecordSchema = z ( reply ): - | (WorktreeDisplayNameSource & { - path?: string - hostId?: string - branch?: string | null - diffComments?: unknown - }) + | (WorktreeDisplayNameSource & { path?: string; hostId?: string; diffComments?: unknown }) | undefined => reply.worktree ) diff --git a/mobile/src/worktree/use-live-worktree-name.ts b/mobile/src/worktree/use-live-worktree-name.ts index 3a09dcb5650..4501b3a1922 100644 --- a/mobile/src/worktree/use-live-worktree-name.ts +++ b/mobile/src/worktree/use-live-worktree-name.ts @@ -30,8 +30,6 @@ export type LiveWorktreeName = { /** Where the worktree lives on its host; null until the host's record arrives. */ path: string | null hostId: string | null - /** The checked-out branch without refs/heads/; null for folders, detached heads and until the record arrives. */ - branch: string | null resolution: WorktreeShowResolution } @@ -50,7 +48,6 @@ export function useLiveWorktreeName({ name: string path?: string | null hostId?: string | null - branch?: string | null }>(() => ({ worktreeId, name: routeNameHint })) // Key verdicts by route so a previous workspace’s deletion cannot escape into the new one. const [resolved, setResolved] = useState<{ @@ -128,16 +125,13 @@ export function useLiveWorktreeName({ if (liveName) { const path = worktree?.path || null const hostId = worktree?.hostId ?? null - // Match desktop formatBranchName by hiding the refs/heads/ prefix. - const branch = worktree?.branch?.replace(/^refs\/heads\//, '') || null setWorktreeName((current) => current.worktreeId === worktreeId && current.name === liveName && current.path === path && - current.hostId === hostId && - current.branch === branch + current.hostId === hostId ? current - : { worktreeId, name: liveName, path, hostId, branch } + : { worktreeId, name: liveName, path, hostId } ) } hasSuccessfulRefresh = true @@ -203,19 +197,12 @@ export function useLiveWorktreeName({ ) if (isFloatingWorkspace) { - return { - name: getFloatingWorkspaceTitle(), - path: null, - hostId: 'local', - branch: null, - resolution: 'unknown' - } + return { name: getFloatingWorkspaceTitle(), path: null, hostId: 'local', resolution: 'unknown' } } return { name: worktreeName.worktreeId === worktreeId ? worktreeName.name : routeNameHint, path: worktreeName.worktreeId === worktreeId ? (worktreeName.path ?? null) : null, hostId: worktreeName.worktreeId === worktreeId ? (worktreeName.hostId ?? null) : null, - branch: worktreeName.worktreeId === worktreeId ? (worktreeName.branch ?? null) : null, resolution: resolved.worktreeId === worktreeId ? resolved.resolution : 'unknown' } } From 4a31139c50b2908a4284940890c9ff92e9625d5a Mon Sep 17 00:00:00 2001 From: Pr1p Date: Wed, 11 Nov 2026 08:46:03 +0800 Subject: [PATCH 09/21] fix(pty): scope Linux terminal foreground rows to the anchor session (#39557) * fix(pty): scope terminal foreground rows to the anchor session on Linux A docker run -t container holds its own private pts/N, and Linux reports terminal membership only by number, so ps -t pts/N listed the container's processes beside the agent's. The foreground coherence check then saw a tpgid that never matched and refused guarded writes with agent_not_in_foreground (#39488). Prefix the Linux terminal ps with a sid column (macOS ps has no sid keyword) and keep only rows in the anchor process's session; a missing anchor row keeps every row as before. Co-Authored-By: Claude Code * fix(pty): require the anchor PID for terminal rows and test the real ps shape The session filter is the fix, so every caller now passes the anchor. The regression test answers in whatever columns ps was asked for, so reverting the fix fails on the foreground coherence check rather than on a parse quirk. --------- Co-authored-by: Claude Code Co-authored-by: Kelvin Amoaba <97001695+AmoabaKelvin@users.noreply.github.com> --- .../pty-master-foreground-observation.ts | 2 +- .../terminal-process-rows.ts | 43 ++++++++++++++-- .../terminal-foreground-capture.test.ts | 51 +++++++++++++++++-- 3 files changed, 89 insertions(+), 7 deletions(-) diff --git a/src/main/pty/process-retirement/pty-master-foreground-observation.ts b/src/main/pty/process-retirement/pty-master-foreground-observation.ts index 70eeb61b08a..f2114a14a1b 100644 --- a/src/main/pty/process-retirement/pty-master-foreground-observation.ts +++ b/src/main/pty/process-retirement/pty-master-foreground-observation.ts @@ -137,7 +137,7 @@ export async function observePtyForeground( ) { return unknown('terminal_unavailable') } - const rows = await readPtyTerminalRows(tty, deadlineMs - Date.now()) + const rows = await readPtyTerminalRows(tty, proc.pid, deadlineMs - Date.now()) if (!alive()) { return unknown('session_changed') } diff --git a/src/main/pty/process-retirement/terminal-process-rows.ts b/src/main/pty/process-retirement/terminal-process-rows.ts index 9dbbf66acd7..60d88fa4d27 100644 --- a/src/main/pty/process-retirement/terminal-process-rows.ts +++ b/src/main/pty/process-retirement/terminal-process-rows.ts @@ -83,20 +83,57 @@ export async function readTerminalTty( return output === null ? null : normalizePtyTty(output) } +/** A container's private pts/N reuses the host's terminal number while its processes keep their own session, so + * ps -t lists them beside the terminal's own (#39488). Strip the sid column ps prefixes and keep only the anchor + * process's session; a missing anchor row keeps every row. */ +function parseTerminalSessionRows(stdout: string, anchorPid: number): ProcessTableRow[] { + const sidByPid = new Map() + const body = stdout + .split(/\r?\n/) + .map((line) => { + const match = line.trim().match(/^(\d+)\s+(\d+)\s+(.*)$/) + if (!match) { + return line.trim() + } + sidByPid.set(Number(match[2]), Number(match[1])) + return `${match[2]} ${match[3]}` + }) + .join('\n') + const rows = parseShellForegroundRows(body, { strict: true }) + const anchorSid = sidByPid.get(anchorPid) + return anchorSid === undefined ? rows : rows.filter((row) => sidByPid.get(row.pid) === anchorSid) +} + export async function readPtyTerminalRows( tty: string, + anchorPid: number, timeoutMs = PS_TIMEOUT_MS ): Promise { const normalized = normalizePtyTty(tty) if (!normalized || timeoutMs <= 0) { return null } + // macOS ps has no sid keyword. + const withSid = process.platform === 'linux' const stdout = await readTerminalPs( - ['-ww', '-o', 'pid=,ppid=,pgid=,tpgid=,stat=,command=', '-t', normalized], + [ + '-ww', + '-o', + withSid + ? 'sid=,pid=,ppid=,pgid=,tpgid=,stat=,command=' + : 'pid=,ppid=,pgid=,tpgid=,stat=,command=', + '-t', + normalized + ], timeoutMs ) try { - return stdout === null ? null : parseShellForegroundRows(stdout, { strict: true }) + if (stdout === null) { + return null + } + return withSid + ? parseTerminalSessionRows(stdout, anchorPid) + : parseShellForegroundRows(stdout, { strict: true }) } catch { return null } @@ -106,7 +143,7 @@ export async function readPtyTerminalRows( export async function readTerminalProcessRows(rootPid: number): Promise { const deadline = Date.now() + PS_TIMEOUT_MS const tty = await readTerminalTty(rootPid, deadline - Date.now()) - const rows = tty ? await readPtyTerminalRows(tty, deadline - Date.now()) : null + const rows = tty ? await readPtyTerminalRows(tty, rootPid, deadline - Date.now()) : null if (!rows?.some((row) => row.pid === rootPid) || Date.now() >= deadline) { return null } diff --git a/src/main/runtime/terminal/lifecycle/terminal-foreground-capture.test.ts b/src/main/runtime/terminal/lifecycle/terminal-foreground-capture.test.ts index 96d609c3d15..1a6c7865845 100644 --- a/src/main/runtime/terminal/lifecycle/terminal-foreground-capture.test.ts +++ b/src/main/runtime/terminal/lifecycle/terminal-foreground-capture.test.ts @@ -1,7 +1,10 @@ import { beforeEach, describe, expect, it, vi } from 'vitest' import { judgeTerminalForeground } from './terminal-foreground-group' +import { + observePtyForeground, + resolvePtyForegroundMembers +} from '../../../pty/process-retirement/pty-master-foreground-observation' import { readTerminalProcessRows } from '../../../pty/process-retirement/terminal-process-rows' -import { resolvePtyForegroundMembers } from '../../../pty/process-retirement/pty-master-foreground-observation' import { readLaunchedAgentForeground } from '../wait/launched-agent-foreground' import { createAdapter } from '../../../daemon/providers/routing/testing/daemon-pty-router-test-fixture' import { DaemonPtyRouter } from '../../../daemon/providers/routing/daemon-pty-router' @@ -10,8 +13,12 @@ const mocks = vi.hoisted(() => ({ runProcess: vi.fn() })) vi.mock('@orca/process-host', () => ({ runProcess: mocks.runProcess })) vi.mock('../../../../shared/libproc-process-table', () => ({ runPsInProcess: async () => null })) -const rows = (pid: number) => +// Linux ps prefixes the sid column readPtyTerminalRows asks for there. +const rows = (pid: number, withSid = process.platform === 'linux') => `${pid} 1 ${pid} ${pid + 1} Ss -zsh\n${pid + 1} ${pid} ${pid + 1} ${pid + 1} S+ claude\n` + .split('\n') + .map((line) => (withSid && line ? `${pid} ${line}` : line)) + .join('\n') const result = (stdout: string, outputTruncated = false) => ({ code: 0, signal: null, @@ -87,9 +94,10 @@ describe('fresh target foreground regressions', () => { expect(await readTerminalProcessRows(90300)).toBeNull() }) it('keeps a capture whose blank-argv row sits outside the foreground group', async () => { + const blankOutside = `${process.platform === 'linux' ? '90400 ' : ''}90410 90400 90410 90401 S` mocks.runProcess .mockResolvedValueOnce(result('ttys904')) - .mockResolvedValueOnce(result(`${rows(90400)}90410 90400 90410 90401 S \n`)) + .mockResolvedValueOnce(result(`${rows(90400)}${blankOutside}\n`)) .mockResolvedValueOnce(result('ttys904')) const captured = (await readTerminalProcessRows(90400)) ?? [] expect(resolvePtyForegroundMembers(captured, 90401).verdict).toBe('observed') @@ -99,3 +107,40 @@ describe('fresh target foreground regressions', () => { expect(resolvePtyForegroundMembers(blank, 90401).verdict).toBe('unverifiable') }) }) + +// #39488: a `docker run -t` container holds its own private pts/0, so host `ps -t pts/0` lists its sleep beside +// the terminal's shell and agent. Rows are the reporter's capture; columns follow what ps was asked for. +const CONTAINER_SHARING_PTS = [ + { sid: 58769, row: '58769 58768 58769 58769 Ss+ sleep 900' }, + { sid: 58824, row: '58824 100 58824 58826 Ss+ /bin/sh -l' }, + { sid: 58824, row: '58826 58824 58826 58826 S+ claude --print-output' } +] + +describe('terminal foreground with a container holding the same pts number', () => { + it('observes the agent as foreground on Linux', async () => { + mocks.runProcess.mockReset() + const original = process.platform + Object.defineProperty(process, 'platform', { configurable: true, value: 'linux' }) + try { + mocks.runProcess.mockImplementation(async ({ args }: { args: readonly string[] }) => { + const withSid = args.some((arg) => arg.startsWith('sid=')) + return result( + `${CONTAINER_SHARING_PTS.map(({ sid, row }) => (withSid ? `${sid} ${row}` : row)).join('\n')}\n` + ) + }) + const owner = { + pid: 58824, + fd: 9, + ptsName: '/dev/pts/0', + getForegroundProcessGroup: (): number | null | undefined => 58826 + } + await expect(observePtyForeground(owner, () => true)).resolves.toMatchObject({ + verdict: 'observed', + foregroundKind: 'non-shell', + foregroundPgid: 58826 + }) + } finally { + Object.defineProperty(process, 'platform', { configurable: true, value: original }) + } + }) +}) From 8ac9fb6eee05fe6d738984aeed7f84aa396789cc Mon Sep 17 00:00:00 2001 From: Brennan Benson <79079362+brennanb2025@users.noreply.github.com> Date: Tue, 10 Nov 2026 16:46:16 -0800 Subject: [PATCH 10/21] Stop re-reading every linked GitHub issue when the window returns (#40534) * fix(github): stop re-reading every linked issue when the window returns Each window reveal re-read the linked GitHub issue of every workspace, shown or not, through the shared 4-slot gh queue, so PR, Checks and Merge waited behind hundreds of issue reads, and above 500 linked issues the cache thrashed on every reveal. Surfaces that show an issue now own its freshness: the reveal handler no longer sweeps the catalog; workspace cards keep their own visible-only issue poll, now also in the paired web client and for folder workspaces; the space manager's decision hover polls its one issue while open. Co-Authored-By: Claude Opus 5.5 (1M context) * fix(space): the open space-manager hover asks the main window to refresh its issue The desktop Space window shows the main window's published issue cache, so a read done in the Space window was overwritten by the next snapshot. The decision hover now owns its issue's freshness only while open, follows the live link, and asks the main window (the cache owner) to re-read it; the web client reads in place. Co-Authored-By: Claude Opus 5.5 (1M context) * fix(space): route the hover's issue refresh by window and follow owner changes Choose the main-window relay only in the separate Space window (the paired web client's fallback API made the relay look available), and refresh when the worktree's owning host changes, not only its link. Co-Authored-By: Claude Opus 5.5 (1M context) * perf(space): key the open hover's issue refresh off the row projection The hover re-resolved its worktree owner from the store on every Space-window snapshot, rebuilding the catalog lookup index each time. Use the cache key the row projection already computes for the issue it shows. Co-Authored-By: Claude Opus 5.5 (1M context) * test(space): use typed fixtures in the hover freshness test Co-Authored-By: Claude Opus 5.5 (1M context) --------- Co-authored-by: Claude Opus 5.5 (1M context) --- .../workspace-space/workspace-space-window.ts | 22 ++++++ src/preload/api/workspace-cleanup-api.ts | 4 + src/preload/api/workspace-space-bridge.ts | 13 ++++ .../use-worktree-card-lifecycle-effects.ts | 16 +--- .../workspace-space-decision-details.ts | 26 ++++--- .../workspace-space-decision-hover-card.tsx | 6 ++ .../workspace-space-issue-freshness.test.tsx | 74 +++++++++++++++++++ .../workspace-space-issue-freshness.tsx | 48 ++++++++++++ .../SpaceWindowStatePublisher.tsx | 5 ++ .../github/refresh/refresh-sweep-actions.ts | 39 +--------- 10 files changed, 191 insertions(+), 62 deletions(-) create mode 100644 src/renderer/src/components/status-bar/workspace-space-issue-freshness.test.tsx create mode 100644 src/renderer/src/components/status-bar/workspace-space-issue-freshness.tsx diff --git a/src/main/ipc/workspace-space/workspace-space-window.ts b/src/main/ipc/workspace-space/workspace-space-window.ts index cd00a4acf48..1c5443a91be 100644 --- a/src/main/ipc/workspace-space/workspace-space-window.ts +++ b/src/main/ipc/workspace-space/workspace-space-window.ts @@ -18,6 +18,7 @@ export function registerWorkspaceSpaceWindowHandlers(store: Store): void { 'workspaceSpace:requestWindowState', 'workspaceSpace:revealWorktree', 'workspaceSpace:requestWorktreeRefresh', + 'workspaceSpace:requestIssueRefresh', 'workspaceSpace:publishWindowAnalysis' ]) { ipcMain.removeHandler(channel) @@ -72,6 +73,27 @@ export function registerWorkspaceSpaceWindowHandlers(store: Store): void { getTrustedUIRendererWebContents()?.send('workspaceSpace:worktreeRefreshRequested') } }) + // Why: the main window owns issueCache and publishes it to the Space window; Space asks rather than writing its own copy. + ipcMain.handle( + 'workspaceSpace:requestIssueRefresh', + (event, worktreeId: unknown, hostId: unknown): void => { + if (!isWorkspaceSpaceWindowRenderer(event.sender) || typeof worktreeId !== 'string') { + return + } + if (hostId !== undefined && typeof hostId !== 'string') { + return + } + const executionHostId = normalizeExecutionHostId(hostId) + if (hostId !== undefined && executionHostId === null) { + return + } + getTrustedUIRendererWebContents()?.send( + 'workspaceSpace:issueRefreshRequested', + worktreeId, + executionHostId ?? undefined + ) + } + ) ipcMain.handle('workspaceSpace:publishWindowAnalysis', (event, analysis: unknown): void => { if (isWorkspaceSpaceWindowRenderer(event.sender) && analysis && typeof analysis === 'object') { getTrustedUIRendererWebContents()?.send('workspaceSpace:windowAnalysis', analysis) diff --git a/src/preload/api/workspace-cleanup-api.ts b/src/preload/api/workspace-cleanup-api.ts index a88023d8e48..84458b41907 100644 --- a/src/preload/api/workspace-cleanup-api.ts +++ b/src/preload/api/workspace-cleanup-api.ts @@ -55,6 +55,10 @@ export type WorkspaceSpaceApi = { ) => () => void requestWorktreeRefresh?: () => Promise onWorktreeRefreshRequested?: (callback: () => void) => () => void + requestIssueRefresh?: (worktreeId: string, executionHostId?: ExecutionHostId) => Promise + onIssueRefreshRequested?: ( + callback: (worktreeId: string, executionHostId?: ExecutionHostId) => void + ) => () => void publishWindowAnalysis?: (analysis: WorkspaceSpaceAnalysis) => Promise onWindowAnalysis?: (callback: (analysis: WorkspaceSpaceAnalysis) => void) => () => void } diff --git a/src/preload/api/workspace-space-bridge.ts b/src/preload/api/workspace-space-bridge.ts index cc222ec5bc1..cfab82ca3bd 100644 --- a/src/preload/api/workspace-space-bridge.ts +++ b/src/preload/api/workspace-space-bridge.ts @@ -21,6 +21,8 @@ export const workspaceSpaceApi = { change?: WorkspaceSpaceChange ) => ipcRenderer.invoke('workspaceSpace:revealWorktree', worktreeId, executionHostId, change), requestWorktreeRefresh: () => ipcRenderer.invoke('workspaceSpace:requestWorktreeRefresh'), + requestIssueRefresh: (worktreeId: string, executionHostId?: ExecutionHostId) => + ipcRenderer.invoke('workspaceSpace:requestIssueRefresh', worktreeId, executionHostId), publishWindowAnalysis: (analysis: WorkspaceSpaceAnalysis) => ipcRenderer.invoke('workspaceSpace:publishWindowAnalysis', analysis), onWindowAnalysis: (callback: (analysis: WorkspaceSpaceAnalysis) => void) => { @@ -65,6 +67,17 @@ export const workspaceSpaceApi = { ipcRenderer.on('workspaceSpace:worktreeRefreshRequested', listener) return () => ipcRenderer.removeListener('workspaceSpace:worktreeRefreshRequested', listener) }, + onIssueRefreshRequested: ( + callback: (worktreeId: string, executionHostId?: ExecutionHostId) => void + ) => { + const listener = ( + _event: Electron.IpcRendererEvent, + worktreeId: string, + executionHostId?: ExecutionHostId + ): void => callback(worktreeId, executionHostId) + ipcRenderer.on('workspaceSpace:issueRefreshRequested', listener) + return () => ipcRenderer.removeListener('workspaceSpace:issueRefreshRequested', listener) + }, onProgress: (callback) => { const listener = ( _event: Electron.IpcRendererEvent, diff --git a/src/renderer/src/components/sidebar/workspace-card/controller/use-worktree-card-lifecycle-effects.ts b/src/renderer/src/components/sidebar/workspace-card/controller/use-worktree-card-lifecycle-effects.ts index f7b381323e8..124e9fc320c 100644 --- a/src/renderer/src/components/sidebar/workspace-card/controller/use-worktree-card-lifecycle-effects.ts +++ b/src/renderer/src/components/sidebar/workspace-card/controller/use-worktree-card-lifecycle-effects.ts @@ -114,16 +114,9 @@ export function useWorktreeCardLifecycleEffects({ ]) // Poll issues only while visible; hidden metadata does not justify GitHub calls. + // Why web clients and folders too: this card is the only owner of its issue's freshness; nothing sweeps the catalog. useEffect(() => { - // Why: per-card decoration lookups from the browser flood the RPC path at paired-web startup; the host is authoritative. - if ( - isWebClient() || - !repo || - isFolder || - !worktree.linkedIssue || - !issueCacheKey || - !showIssue - ) { + if (!repo || !worktree.linkedIssue || !issueCacheKey || !showIssue) { return } @@ -134,16 +127,14 @@ export function useWorktreeCardLifecycleEffects({ run: () => void fetchIssue(repo.path, issueNumber, { repoId: repo.id }), intervalMs: 5 * 60_000 }) - }, [repo, isFolder, worktree.linkedIssue, fetchIssue, issueCacheKey, showIssue]) + }, [repo, worktree.linkedIssue, fetchIssue, issueCacheKey, showIssue]) useEffect(() => { if ( !newCardStyle || !hoverDetailsOpen || showIssue || - isWebClient() || !repo || - isFolder || !worktree.linkedIssue || !issueCacheKey ) { @@ -155,7 +146,6 @@ export function useWorktreeCardLifecycleEffects({ hoverDetailsOpen, showIssue, repo, - isFolder, worktree.linkedIssue, fetchIssue, issueCacheKey diff --git a/src/renderer/src/components/status-bar/workspace-space-decision-details.ts b/src/renderer/src/components/status-bar/workspace-space-decision-details.ts index aa44b91f341..f135c259d55 100644 --- a/src/renderer/src/components/status-bar/workspace-space-decision-details.ts +++ b/src/renderer/src/components/status-bar/workspace-space-decision-details.ts @@ -33,6 +33,8 @@ export type WorkspaceDecisionDetails = { branchStatus: string | null reviewLabel: string | null issueLabel: string | null + /** Cache key of the shown linked issue; changes when the link or its owning host changes. */ + issueCacheKey: string | null linearIssueLabel: string | null } @@ -169,20 +171,19 @@ export function getWorkspaceDecisionDetails( ? `PR #${linkedPR}` : null const linkedIssue = workspaceRecord?.linkedIssue ?? null - const issue = + const issueCacheKey = linkedIssue && repo - ? inputs.issueCache[ - getIssueCacheKey( - repo.path, - repo.id, - linkedIssue, - inputs.settings, - repo.connectionId, - ownerExecutionHostId, - true - ) - ]?.data + ? getIssueCacheKey( + repo.path, + repo.id, + linkedIssue, + inputs.settings, + repo.connectionId, + ownerExecutionHostId, + true + ) : null + const issue = issueCacheKey ? inputs.issueCache[issueCacheKey]?.data : null const issueLabel = linkedIssue ? issue ? `#${issue.number} ${issue.state}: ${issue.title}` @@ -228,6 +229,7 @@ export function getWorkspaceDecisionDetails( branchStatus: getBranchStatus(inputs.remoteStatusesByWorktree[worktree.worktreeId], t), reviewLabel, issueLabel, + issueCacheKey, linearIssueLabel } } diff --git a/src/renderer/src/components/status-bar/workspace-space-decision-hover-card.tsx b/src/renderer/src/components/status-bar/workspace-space-decision-hover-card.tsx index a7db8f3d867..63ba315e0a1 100644 --- a/src/renderer/src/components/status-bar/workspace-space-decision-hover-card.tsx +++ b/src/renderer/src/components/status-bar/workspace-space-decision-hover-card.tsx @@ -16,6 +16,7 @@ import { useTranslate } from '@/i18n/use-translate' import type { WorkspaceDecisionDetails } from './workspace-space-decision-details' import type { WorkspaceGitRefreshState } from './workspace-space-manager-state-types' import { WorkspaceSpaceStatusBadge } from './workspace-space-status-badge' +import { WorkspaceSpaceIssueFreshness } from './workspace-space-issue-freshness' import { getAgentDecisionLabel, getDeleteDecisionLabel, @@ -81,6 +82,11 @@ export function WorkspaceDecisionHoverCard({ collisionPadding={12} className="max-h-[min(34rem,calc(100vh-1.5rem))] w-[min(24rem,calc(100vw-1.5rem))] overflow-y-auto p-0 scrollbar-sleek" > +
diff --git a/src/renderer/src/components/status-bar/workspace-space-issue-freshness.test.tsx b/src/renderer/src/components/status-bar/workspace-space-issue-freshness.test.tsx new file mode 100644 index 00000000000..6a3cbf84542 --- /dev/null +++ b/src/renderer/src/components/status-bar/workspace-space-issue-freshness.test.tsx @@ -0,0 +1,74 @@ +// @vitest-environment happy-dom + +import { act } from 'react' +import { createRoot, type Root } from 'react-dom/client' +import { afterEach, beforeEach, expect, it, vi } from 'vitest' +import { useAppStore } from '@/store' +import type { AppState } from '@/store/types' +import { WorkspaceSpaceIssueFreshness } from './workspace-space-issue-freshness' +import { makeRepo, makeWorktree } from '../worktree-jump/worktree-jump-palette-test-fixtures' + +let container: HTMLDivElement +let root: Root +const fetchIssue = vi.fn(async () => null) + +function linkIssue(linkedIssue: number | null): void { + useAppStore.setState({ + repos: [makeRepo()], + worktreesByRepo: { 'repo-1': [makeWorktree('repo-1::/wt/a', 'a', { linkedIssue })] }, + fetchIssue + }) +} + +beforeEach(() => { + useAppStore.setState(useAppStore.getInitialState(), true) + fetchIssue.mockClear() + container = document.createElement('div') + document.body.appendChild(container) + root = createRoot(container) +}) + +afterEach(() => { + act(() => root.unmount()) + container.remove() + Reflect.deleteProperty(window, 'api') + window.history.replaceState(null, '', '/') +}) + +it('reads the live linked issue while open, following a late and a changed link', () => { + const requestIssueRefresh = vi.fn(async () => {}) + Object.assign(window, { api: { workspaceSpace: { requestIssueRefresh } } }) + const show = (issueCacheKey: string | null): void => + act(() => + root.render( + + ) + ) + linkIssue(null) + show(null) + expect(fetchIssue).not.toHaveBeenCalled() + + linkIssue(91) + show('repo-1::91') + expect(fetchIssue).toHaveBeenLastCalledWith('/repos/repo-1', 91, { repoId: 'repo-1' }) + + linkIssue(92) + show('repo-1::92') + expect(fetchIssue).toHaveBeenLastCalledWith('/repos/repo-1', 92, { repoId: 'repo-1' }) + expect(requestIssueRefresh).not.toHaveBeenCalled() +}) + +it('asks the main window, which owns the issue cache, from the desktop Space window', () => { + const requestIssueRefresh = vi.fn(async () => {}) + Object.assign(window, { api: { workspaceSpace: { requestIssueRefresh } } }) + window.history.replaceState(null, '', '/space.html') + linkIssue(91) + act(() => + root.render( + + ) + ) + + expect(requestIssueRefresh).toHaveBeenCalledWith('repo-1::/wt/a', undefined) + expect(fetchIssue).not.toHaveBeenCalled() +}) diff --git a/src/renderer/src/components/status-bar/workspace-space-issue-freshness.tsx b/src/renderer/src/components/status-bar/workspace-space-issue-freshness.tsx new file mode 100644 index 00000000000..004742c7f9b --- /dev/null +++ b/src/renderer/src/components/status-bar/workspace-space-issue-freshness.tsx @@ -0,0 +1,48 @@ +import { useEffect } from 'react' +import type { ExecutionHostId } from '../../../../shared/execution-host' +import { useAppStore } from '@/store' +import { installWindowVisibilityInterval } from '@/lib/window-visibility-interval' +import { findWorktreeReviewOwner } from '@/store/github/refresh/worktree-refresh' + +/** Re-reads the worktree's current linked issue in this window's store; fresh-gated by fetchIssue. */ +export function refreshLinkedIssueForWorktree( + worktreeId: string, + executionHostId?: ExecutionHostId +): void { + const state = useAppStore.getState() + const owner = findWorktreeReviewOwner(state, worktreeId, executionHostId) + const issueNumber = owner?.worktree.linkedIssue + if (owner && issueNumber) { + void state.fetchIssue(owner.repo.path, issueNumber, { repoId: owner.repo.id }) + } +} + +/** Mounted only inside an open decision hover: that hover owns its linked issue's freshness. */ +export function WorkspaceSpaceIssueFreshness({ + worktreeId, + executionHostId, + issueCacheKey +}: { + worktreeId: string + executionHostId?: ExecutionHostId + // Why a prop: the row projection already resolves it per snapshot; re-resolving here would rescan the catalog. + issueCacheKey: string | null +}): null { + useEffect(() => { + if (!issueCacheKey) { + return + } + // Why: the separate Space window shows the main window's published issueCache, so the main window does the read. + const isSpaceWindow = window.location.pathname.endsWith('/space.html') + return installWindowVisibilityInterval({ + run: () => + isSpaceWindow + ? void window.api.workspaceSpace + .requestIssueRefresh?.(worktreeId, executionHostId) + .catch(() => {}) + : refreshLinkedIssueForWorktree(worktreeId, executionHostId), + intervalMs: 5 * 60_000 + }) + }, [issueCacheKey, worktreeId, executionHostId]) + return null +} diff --git a/src/renderer/src/components/workspace-space/SpaceWindowStatePublisher.tsx b/src/renderer/src/components/workspace-space/SpaceWindowStatePublisher.tsx index 9d9939c5540..5059a88447d 100644 --- a/src/renderer/src/components/workspace-space/SpaceWindowStatePublisher.tsx +++ b/src/renderer/src/components/workspace-space/SpaceWindowStatePublisher.tsx @@ -7,6 +7,7 @@ import { type SpaceWindowState } from './space-window-state' import { revealWorkspaceSpaceChange } from '../status-bar/workspace-space-open-change' +import { refreshLinkedIssueForWorktree } from '../status-bar/workspace-space-issue-freshness' export function SpaceWindowStatePublisher(): null { useEffect(() => { @@ -70,6 +71,9 @@ export function SpaceWindowStatePublisher(): null { const offRefresh = api.onWorktreeRefreshRequested?.(() => { void useAppStore.getState().fetchAllWorktrees() }) + const offIssueRefresh = api.onIssueRefreshRequested?.((worktreeId, executionHostId) => { + refreshLinkedIssueForWorktree(worktreeId, executionHostId) + }) const offAnalysis = api.onWindowAnalysis?.((analysis) => { if (typeof analysis.scannedAt === 'number') { useAppStore.setState({ workspaceSpaceAnalysis: analysis }) @@ -84,6 +88,7 @@ export function SpaceWindowStatePublisher(): null { offClose?.() offReveal?.() offRefresh?.() + offIssueRefresh?.() offAnalysis?.() } }, []) diff --git a/src/renderer/src/store/github/refresh/refresh-sweep-actions.ts b/src/renderer/src/store/github/refresh/refresh-sweep-actions.ts index 48ea4b3a385..6acd200b589 100644 --- a/src/renderer/src/store/github/refresh/refresh-sweep-actions.ts +++ b/src/renderer/src/store/github/refresh/refresh-sweep-actions.ts @@ -3,13 +3,11 @@ import type { AppState } from '../../types' import type { GitHubSlice } from '../github-slice-contract' import { getRepoExecutionHostId } from '../../../../../shared/execution-host' import { buildPRRefreshCandidate, findWorktreeReviewOwner } from './worktree-refresh' -import { getGitHubRepoLookupIndex } from '../repositories/github-repo-lookup-index' import { issueCacheKey } from '../cache/cache-identity' -import { CACHE_TTL, evictStaleEntries } from '../cache/cache-policy' +import { evictStaleEntries } from '../cache/cache-policy' import { pruneExpiredPRRefreshStates } from './pr-refresh-state' import { pruneExpiredPRRefreshStartedEntries } from '../pull-requests/pr-result-cache' import { settingsForGitHubRepoOwner } from '../work-items/work-item-routing' -import { selectWorktreesByRepo } from '../../workspaces/catalog/workspace-record-fields' export const createRefreshSweepActions = ( set: Parameters>[0], @@ -41,40 +39,7 @@ export const createRefreshSweepActions = ( }) // Why: don't prune prRequestGenerations here — deleting a live generation makes its response look stale. - - const state = get() - if (!(state.worktreeCardProperties ?? []).includes('issue')) { - return - } - - const now = Date.now() - const repoLookup = getGitHubRepoLookupIndex(state.repos) - - for (const worktrees of Object.values(selectWorktreesByRepo(state))) { - for (const wt of worktrees) { - const repo = repoLookup.findById(wt.repoId) - if (!repo) { - continue - } - - if (wt.linkedIssue) { - const ownerSettings = settingsForGitHubRepoOwner(state.settings, repo) - const issueKey = issueCacheKey( - repo.path, - repo.id, - wt.linkedIssue, - ownerSettings, - repo.connectionId, - repo.executionHostId, - true - ) - const issueEntry = state.issueCache[issueKey] - if (!issueEntry || now - issueEntry.fetchedAt >= CACHE_TTL) { - void get().fetchIssue(repo.path, wt.linkedIssue, { repoId: repo.id }) - } - } - } - } + // Each surface that shows a linked issue refreshes it on reveal; no catalog-wide issue reads here. }, refreshGitHubForWorktree: (worktreeId, executionHostId) => { From 049c0a74218fed84fadb0343c79e751b58319ce7 Mon Sep 17 00:00:00 2001 From: OrcaWin Date: Tue, 10 Nov 2026 16:52:02 -0800 Subject: [PATCH 11/21] test: remove eight low-value suites and routine renderer cases (#40941) Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com> --- config/reliability-gates.jsonc | 79 ++--- .../scripts/ci/sharding/ci-shard-timings.json | 6 - .../WorktreeContextMenuDeleteItems.test.tsx | 78 ----- .../deletion/active-workspace-delete.test.ts | 283 ------------------ ...ty-transport-healthy-link-reattach.test.ts | 56 ---- ...y-transport-unpublished-host-graph.test.ts | 30 -- .../ipc-events/runtime-file-position.test.ts | 74 ----- ...EditorExternalWatch-subscriptions.test.tsx | 47 --- ...ntime-file-write-disk-precondition.test.ts | 99 ------ ...runtime-file-client-search-listing.test.ts | 106 ------- ...n-terminal-pending-handle-recovery.test.ts | 149 --------- .../remote-runtime-snapshot-outcome.test.ts | 13 - .../ssh/runtime-environment-ssh-state.test.ts | 50 ---- ...ructured-agent-session-status-feed.test.ts | 128 -------- ...n-tabs-sync-cross-runtime-worktree.test.ts | 91 +----- ...ssion-open-files-batch-equivalence.test.ts | 212 ------------- ...moved-frame-retires-live-publisher.test.ts | 46 --- .../server/server-held-panes.test.tsx | 21 -- .../provider-request-coalescing.test.ts | 51 ---- .../store-session-editor-hydration.test.ts | 272 ----------------- src/renderer/src/store/slices/ssh/ssh.test.ts | 214 ------------- .../worktrees-terminal-pr-url-linking.test.ts | 149 --------- .../worktrees-purge-terminal-state.test.ts | 241 --------------- .../store-session-workspace-hydration.test.ts | 250 ---------------- ...rminal-pane-detach-agent-retention.test.ts | 240 --------------- 25 files changed, 34 insertions(+), 2951 deletions(-) delete mode 100644 src/renderer/src/components/sidebar/workspace-actions/context-menu/WorktreeContextMenuDeleteItems.test.tsx delete mode 100644 src/renderer/src/components/sidebar/workspace-actions/deletion/active-workspace-delete.test.ts delete mode 100644 src/renderer/src/runtime/web-session-tabs-sync/placement/web-session-open-files-batch-equivalence.test.ts delete mode 100644 src/renderer/src/store/slices/editor/hydration/store-session-editor-hydration.test.ts delete mode 100644 src/renderer/src/store/slices/ssh/ssh.test.ts delete mode 100644 src/renderer/src/store/slices/worktrees/teardown/worktrees-purge-terminal-state.test.ts delete mode 100644 src/renderer/src/store/terminals/store-session-workspace-hydration.test.ts delete mode 100644 src/renderer/src/store/terminals/terminal-pane-detach-agent-retention.test.ts diff --git a/config/reliability-gates.jsonc b/config/reliability-gates.jsonc index 0799cf6b242..f992e3fbfae 100644 --- a/config/reliability-gates.jsonc +++ b/config/reliability-gates.jsonc @@ -4186,10 +4186,10 @@ "providers": ["remote-runtime"], "coveredPlatforms": ["macos"], "coveredProviders": ["remote-runtime"], - "coverageNotes": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Deterministic main-IPC contract tests cover disconnect-driven close delivery, exactly-once close, per-subscription teardown isolation against a failing socket close and a throwing liveness probe, containment of a throwing renderer send on the unguarded host-close path, and continued suppression of stale payloads from a retired transport. A headed paired-server journey (real Orca host plus a separate paired Orca desktop client) covers hidden-but-mounted reveal, cold-parked reveal, and cold-parked reveal across a disconnect/reconnect. Live Linux and Windows paired-server evidence and real sleep/wake transport loss remain uncollected.", + "coverageNotes": "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Deterministic main-IPC contract tests cover disconnect-driven close delivery, exactly-once close, per-subscription teardown isolation against a failing socket close and a throwing liveness probe, containment of a throwing renderer send on the unguarded host-close path, and continued suppression of stale payloads from a retired transport. A headed paired-server journey (real Orca host plus a separate paired Orca desktop client) covers hidden-but-mounted reveal, cold-parked reveal, and cold-parked reveal across a disconnect/reconnect. Live Linux and Windows paired-server evidence and real sleep/wake transport loss remain uncollected.", "motivatingLinks": ["tests/e2e/paired-remote-terminal-parked-reveal-interactivity.spec.ts"], "invariant": "Every renderer-held runtime subscription receives exactly one terminal close event when its transport is retired, including when the retirement advanced the transport generation first, and a single failing teardown never abandons that environment's remaining subscriptions nor escapes into the transport that reported the close. Payload frames from a retired transport stay suppressed. A revealed remote terminal therefore reattaches over a live multiplex connection: its buffer restores, typed input reaches the host PTY, the echo paints without a tab flip, and the PTY converges on the revealed pane grid.", - "oracle": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): The main IPC contract test subscribes terminal.multiplex through the real handler, disconnects the environment, and asserts the renderer received exactly one {type: close} subscription event. Two isolation tests subscribe a second stream to the same environment and make the first one fail -- in its socket close, and in the liveness probe inside notifyClosed -- then assert the disconnect does not throw, both transports closed, and every close the renderer could still receive was delivered. A third drives a host-initiated close through the transport callback, which is the one notifyClosed call site with no surrounding guard, with a renderer send that throws, and asserts it cannot escape into the WebSocket close handler. A fourth test asserts that after retirement a late response frame is not forwarded and a late transport close does not re-send. The paired-server journey runs three reveal scenarios against one real host and one real paired desktop client, and for each records buffer restore, host-side receipt of the typed marker through an out-of-band host sink file, live paint without a tab flip, and PTY-versus-pane grid convergence.", + "oracle": "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): The main IPC contract test subscribes terminal.multiplex through the real handler, disconnects the environment, and asserts the renderer received exactly one {type: close} subscription event. Two isolation tests subscribe a second stream to the same environment and make the first one fail -- in its socket close, and in the liveness probe inside notifyClosed -- then assert the disconnect does not throw, both transports closed, and every close the renderer could still receive was delivered. A third drives a host-initiated close through the transport callback, which is the one notifyClosed call site with no surrounding guard, with a renderer send that throws, and asserts it cannot escape into the WebSocket close handler. A fourth test asserts that after retirement a late response frame is not forwarded and a late transport close does not re-send. The paired-server journey runs three reveal scenarios against one real host and one real paired desktop client, and for each records buffer restore, host-side receipt of the typed marker through an out-of-band host sink file, live paint without a tab flip, and PTY-versus-pane grid convergence.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts", "pnpm exec playwright test tests/e2e/paired-remote-terminal-parked-reveal-interactivity.spec.ts --config tests/playwright.config.ts --project electron-headless --workers=1" @@ -4296,7 +4296,8 @@ "Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical.", "Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed.", "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed.", - "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed." + "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed.", + "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed." ], "demotionRule": "Demote if a retired runtime transport can leave a renderer subscription without a close event, if a close is delivered more than once, if one failing teardown strands its sibling subscriptions, if payload frames from a retired transport reach the renderer, or if a revealed remote terminal can stay blank or reject input after a reconnect." }, @@ -8519,15 +8520,15 @@ ], "platforms": ["macos", "linux", "windows"], "providers": ["paired-runtime", "ssh"], - "coveredPlatforms": ["macos", "linux"], + "coveredPlatforms": ["macos"], "coveredProviders": [], - "coverageNotes": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. The direct side-effect-facts selector retains six cases: replay exclusion, live sequence handling, stale disposer ownership, one handoff drain, drain-before-snapshot, and stale replacement-snapshot refusal. Successful current-entry snapshot application is not witnessed by the retained snapshot case. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2162: Reduced inventory-authority source intentionally loses the examples recorded in knownGaps; historical execution is not validation of changed bytes. Historical prior account (preserved verbatim): Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2003: Remove SSH/absent-inventory attach refusal, later-subscriber retry, no-view adoption/takeback, provider-tail/unknown-read projections, failed-keystroke retry and simpler loss reattach examples. Keep the existing-subscriber inventory-ready retry alias only at its retained callback, plus retained held-input cancellation/replacement and overlapping attach/loss fences. The inventory driver also loses the whole empty-worktree-publication registration and folder-host/unverifiable scalar examples. Retire the entire dated10-case attach receipt; retain16 independent receipts only in their original scopes. No retained-memory or real-daemon measurement follows from supplied provider inventories. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1991: Remove initial-snapshot sequence-domain, limited-tail/truncation and unknown-output-sequence aliases whose callbacks are removed. The source retains partial-chunk replay and requested-snapshot retry/admission at supplied snapshot/data seams, not equivalent replacement or measured retained-memory proof. Preserve the existing stream/pending-wait capacity aliases only to their retained owning multiplex callbacks with supplied PTY resolutions. Removing the sibling-PTY heartbeat dedup work-count sample does not establish a new decorative fanout budget; the unrelated-PTY cancellation fence survives. Preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Remove the reasonless-close selector and narrow census aliases to actual retained runtime and injected-inventory callbacks. The removed synthetic census-dedup and isolated pre-boundary timing cases receive no replacement credit; preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1939 (sources1934/1935): Delete the synthetic watcher-synchronization and partial-reconciliation suites. Retire all affected host-fact, capability-owner, activation-residue and retention-cap aliases because the remaining callbacks do not establish those full claims. Keep exact local hook/timer, byte-stream side effects, exit/disposal, scroll-intent cleanup and deterministic local rank/TTL progression callbacks within their mocked-host scope. No physical paired-host retention, resource budget or equivalent replacement evidence is assigned. The generic broad test command remains a runnable discovery command, never a dated execution receipt. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete hook-agent-status projection suite and its two behavioral assertion aliases. Supplied stored hook/title/resume/staleness/rename/question/generation/PTY-retirement/snapshot-version projections are intentionally lost. The separate osc-status-scan gate mentions this source only in existing historical notes and receives no mutation. No measured hidden-output retention, CPU/RSS or resource budget result is claimed by the deleted source or this audit. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Module ownership cleanup retires the modeled cold-park private watcher/render/work-count suites under the current test policy. Their assertions are no longer active evidence. Retained lifecycle tests and headed/headless journeys do not establish equivalent coverage for those removed harness alarms. Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1627 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Retained exact unchanged-source receipts remain the only current execution evidence. Reduced selectors have no fresh execution result. Audit1417 intentionally deletes src/renderer/src/components/terminal-pane/pty-connection-hidden-tui-snapshot-replay.spec.ts: Loses rich hidden snapshot sequence replay/grid resize and forced repaint after hidden snapshot alarms, including live-trigger deduplication. deletes src/renderer/src/components/terminal-pane/terminal-cold-park-verdict-loop.test.tsx: Loses modeled React oscillation alarm; real pane/watchers are not exercised. deletes src/renderer/src/runtime/remote-runtime-terminal-parse-backpressure.test.ts: Loses delayed credit until parsing, malformed frame/unknown stream handling, discarded transformed output recovery and failed ACK rearm alarms. No equivalent replacement coverage or fresh result is claimed. Historical coverage account: Audit1412 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): This audit removes the browser response-after-manual-disconnect example. Its distinct in-flight response fencing alarm is lost; retained queued-call disconnect tests are not an equivalent replacement. Historical commands and results remain historical. Deterministic headed macOS runs launch an isolated Orca desktop server and a separate paired web client. A byte-identical headless run uses an isolated `orca serve` host. Both create six real paired host PTYs with bounded high-output scrollback, prove sustained output while all six client xterms are warm-mounted but hidden causes zero renderer scheduler work, then drive 180 timed real-PTY spinner frames through a second production `session.tabs.subscribeAll` transport and the real Zustand store. Each worktree must publish and apply only its exact initial-working and final-idle titles. Both topologies ordinary-park five xterms without enabling the lossy retention budget, assert bounded cells/heap/timer lag, then restore the bounded authoritative tail exactly once on the original PTY including output produced while parked and continued input/output. The headed oracle additionally keeps every remote target workspace and terminal unmounted in the host renderer before and after client recovery. Production-parity React contracts reproduce and remove the independent cold-park watcher/title feedback loop, including memo-cache discard, StrictMode cleanup replay, detached split-PTY replacement, and the zero-park fast path. A deterministic cross-boundary contract drives 24 real PTY trackers in distinct worktrees through 2,880 decorative spinner and bare-native redraw chunks and counts host publications, serialized bytes, renderer apply calls, and store mutations while preserving exact per-PTY raw chunks and per-worktree semantic transitions. Fake-clock continuation beyond the 30-minute stale boundary requires one deduplicated refresh per worktree every 15 minutes, globally spaced by at least 50 ms, with every mirrored working status still fresh. Exact-identical Pi plus rotating normalized Pi/Grok frames renew coarse working/permission state; rich prompt/tool/question fields retain their original 30-minute lease, and `stateStartedAt`/acknowledgement identity remain stable across title/hook source swaps. A separate headed desktop-client journey relaunches the host main process while retaining the daemon and a daemon-backgrounded PTY, then requires first-observation foreground sync and continued input/output on the same PTY. Unit coverage keeps layout versions out of snapshot output high-waters when a restored snapshot has no output sequence and retries an existing subscriber only after authoritative provider inventory proves the retained PTY exists. Separate headed/headless ACK-starvation tests recover one stalled stream without replacing its PTY. A headed output-drop oracle proves a successful snapshot probe cannot certify a stale live stream when the authoritative PTY sequence advanced beyond the client high-water. Unit coverage requires an unsequenced client to establish a first-probe baseline, remain attached at the same sequence, and recover only after a later probe advances. A capability-disabled run proves legacy hosts retain the prior lossy limit/TTL fallback. Unit tests cover exact-owner capability routing, raw-stream release, singleton side-effect facts with timed handoff cleanup, 64 simultaneous synthetic-title sources with zero decorative client events after convergence, per-client late-subscribe recovery, semantic bell/title transitions, local animation-frame preservation, legacy no-output-pause delivery gating with exact snapshot restore, provider-authoritative snapshots, 128 active streams plus retry after a 129th-stream capacity rejection, capacity-pressure backoff, full split-leaf remint reconciliation, truncation, and manual-disconnect queue fencing. Linux, Windows, live SSH, and production-scale paired hosts remain gaps. The audit removed the E2E parking override/debug-handle suite. Override exposure guards, delay/retention independence, TTL filtering and debug-handle forwarding alarms are lost. Retained real parking journeys do not establish equivalent test-harness configuration coverage. The test-value audit removes separate responsive-stream, intervening-control-frame and capacity-error classification examples. Retained multiplexer checks use fake clocks and captured transport callbacks; their per-stream isolation, deadline and probe-baseline assertions do not prove an actual host restart or equivalent replacement coverage. Historical run commands and results remain historical. The test-value audit removes the standalone retained/live metadata projection and additive-preference negative example from runtime-client-events.test.ts. Its remaining case checks partial-authority event admission through a captured callback; it does not establish full-pair projection, actual peer delivery or equivalent replacement coverage. Historical commands and results remain historical. The isolated accepted-handle-events observer suite is removed. Its runtime/worktree/pane membership, once-only delivery and listener-release ledger alarms are lost. Live hidden-retention evidence does not establish equivalent observer coverage. Audit1292: The entire host-session mirror empty-inventory settle suite is removed. Live-unpublished versus genuinely-empty hydration, same-ID revision response/in-flight fencing and cleared-tracking pending-probe invalidation alarms are lost. The singleton store and inventories were mocked; no real host census or resume delivery or equivalent replacement coverage is claimed. Audit1302: src/main/runtime/session-tabs-inventory-publication.test.ts: Keep actual runtime public inventory authority across renderer epochs/resync, fallback, aborts and invalid graph publishers. Remove copied nonempty publication, private scan counters/collector and known-host replacement, fake refresh that performs the restoration itself, environment capability flag, and private resume-store characterization whose fixture supplies reservation/rollback ordering. Lose isolated refresh/churn/receiver/capability and copied-payload wiring alarms; no equivalent replacement. Reduced selectors require fresh validation; prior full-suite receipts do not validate removed or changed coverage. No equivalent replacement or physical-provider claim. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1318: src/shared/remote-runtime/control/remote-runtime-shared-control-connection.test.ts: Retain real local encrypted WebSocket regressions for pre-ready socket replacement preserving a request, close-before-ready logical subscription cleanup, unrelated RPC surviving another deadline, and pending setup abort leaving a sibling pending. Host server is synthetic and uses production crypto; kept scenarios also retain internal diagnostics/memory assertions unchanged. Remove happy routing, authority-envelope, unknown-ID/log, cleanup enum, binary, keepalive/liveness and production-budget/private socket inventories; intentionally lose those alarms, no physical host authorization, independent crypto or remote provider claim. Reduced selectors need fresh validation; historical full-suite receipts do not validate the changed coverage. No equivalent replacement or physical-provider claim. Audit1346 removes the entire private-state-injected session-tabs rich-status boundary suite. Its stale task/question, neutral-title preservation, owner-change and provider-generation alarms are intentionally lost. Retained real terminal journeys do not establish equivalent replacement coverage. The prior mixed 42-contract execution receipt is historical and must not validate the reduced current selector; fresh validation remains required. Audit1352: focused Linux unit selections validated on source head b638e07d395b76af3ba6b03761a1a0fe7e5e026e. Historical statements above retain their original scope. Intentional test losses are recorded in knownGaps; these runs claim no physical-provider, p95, rendered E2E or cross-platform validation. Audit1360: focused Linux unit selections validated on source head 415c9cffc2d10024131ec4b6721767a22195de6d. Historical statements above retain their original scope. Intentional test losses are recorded in knownGaps; these runs claim no physical-provider, p95, rendered E2E or cross-platform validation. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1391: src/renderer/src/components/terminal-pane/pty-connection-remote-pause-drop-notice.spec.ts reduced. Lose repeated quiet switches avoiding snapshots and legacy host without drop-report capability requesting a snapshot. Retained imported registrar uses fake transport/xterm writes, exercising real gap/restore lifecycle rather than actual terminal rendering. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1396: src/renderer/src/components/terminal-pane/terminal-parked-watcher-reconciliation.test.ts reduced. Retire standalone reconcile set output and positive untouchedFreshSpawn/workspaceOwner copying alarms; no native watcher execution claim. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1402: src/renderer/src/components/terminal-pane/pty-connection-park-reveal-model-snapshot.spec.ts reduced. Loses basic paired serialize/initial replay forwarding check; hidden alt-frame grid/precedence policy remains. No real terminal renderer claim. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1412: src/main/runtime/mobile-session-tabs-agent-status-heartbeat.test.ts reduced. Loses FIFO fairness/spacing, semantic bypass/cancellation, delayed-boundary scheduling and teardown timer-cleanup alarms. Retained fixed emission cases do not replace those outcomes. No equivalent replacement coverage is claimed.", + "coverageNotes": "Audit 3318 reduces remote-runtime-pty-transport-unpublished-host-graph.test.ts, retiring its direct published-empty-graph omission-to-pane-exit alarm. The unchanged bounded unpublished-frame/pushed-recovery and stale-handle park/manual-retry cases are narrower, not equivalent replacements. Both complete three-case Linux receipts dated 2026-11-04 are archived as historical, including source heads; they do not validate this reduced selector. Remaining fourteen active receipts are macOS only. No fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. The direct side-effect-facts selector retains six cases: replay exclusion, live sequence handling, stale disposer ownership, one handoff drain, drain-before-snapshot, and stale replacement-snapshot refusal. Successful current-entry snapshot application is not witnessed by the retained snapshot case. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2162: Reduced inventory-authority source intentionally loses the examples recorded in knownGaps; historical execution is not validation of changed bytes. Historical prior account (preserved verbatim): Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2003: Remove SSH/absent-inventory attach refusal, later-subscriber retry, no-view adoption/takeback, provider-tail/unknown-read projections, failed-keystroke retry and simpler loss reattach examples. Keep the existing-subscriber inventory-ready retry alias only at its retained callback, plus retained held-input cancellation/replacement and overlapping attach/loss fences. The inventory driver also loses the whole empty-worktree-publication registration and folder-host/unverifiable scalar examples. Retire the entire dated10-case attach receipt; retain16 independent receipts only in their original scopes. No retained-memory or real-daemon measurement follows from supplied provider inventories. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1991: Remove initial-snapshot sequence-domain, limited-tail/truncation and unknown-output-sequence aliases whose callbacks are removed. The source retains partial-chunk replay and requested-snapshot retry/admission at supplied snapshot/data seams, not equivalent replacement or measured retained-memory proof. Preserve the existing stream/pending-wait capacity aliases only to their retained owning multiplex callbacks with supplied PTY resolutions. Removing the sibling-PTY heartbeat dedup work-count sample does not establish a new decorative fanout budget; the unrelated-PTY cancellation fence survives. Preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Remove the reasonless-close selector and narrow census aliases to actual retained runtime and injected-inventory callbacks. The removed synthetic census-dedup and isolated pre-boundary timing cases receive no replacement credit; preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1939 (sources1934/1935): Delete the synthetic watcher-synchronization and partial-reconciliation suites. Retire all affected host-fact, capability-owner, activation-residue and retention-cap aliases because the remaining callbacks do not establish those full claims. Keep exact local hook/timer, byte-stream side effects, exit/disposal, scroll-intent cleanup and deterministic local rank/TTL progression callbacks within their mocked-host scope. No physical paired-host retention, resource budget or equivalent replacement evidence is assigned. The generic broad test command remains a runnable discovery command, never a dated execution receipt. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete hook-agent-status projection suite and its two behavioral assertion aliases. Supplied stored hook/title/resume/staleness/rename/question/generation/PTY-retirement/snapshot-version projections are intentionally lost. The separate osc-status-scan gate mentions this source only in existing historical notes and receives no mutation. No measured hidden-output retention, CPU/RSS or resource budget result is claimed by the deleted source or this audit. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Module ownership cleanup retires the modeled cold-park private watcher/render/work-count suites under the current test policy. Their assertions are no longer active evidence. Retained lifecycle tests and headed/headless journeys do not establish equivalent coverage for those removed harness alarms. Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1627 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Retained exact unchanged-source receipts remain the only current execution evidence. Reduced selectors have no fresh execution result. Audit1417 intentionally deletes src/renderer/src/components/terminal-pane/pty-connection-hidden-tui-snapshot-replay.spec.ts: Loses rich hidden snapshot sequence replay/grid resize and forced repaint after hidden snapshot alarms, including live-trigger deduplication. deletes src/renderer/src/components/terminal-pane/terminal-cold-park-verdict-loop.test.tsx: Loses modeled React oscillation alarm; real pane/watchers are not exercised. deletes src/renderer/src/runtime/remote-runtime-terminal-parse-backpressure.test.ts: Loses delayed credit until parsing, malformed frame/unknown stream handling, discarded transformed output recovery and failed ACK rearm alarms. No equivalent replacement coverage or fresh result is claimed. Historical coverage account: Audit1412 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): This audit removes the browser response-after-manual-disconnect example. Its distinct in-flight response fencing alarm is lost; retained queued-call disconnect tests are not an equivalent replacement. Historical commands and results remain historical. Deterministic headed macOS runs launch an isolated Orca desktop server and a separate paired web client. A byte-identical headless run uses an isolated `orca serve` host. Both create six real paired host PTYs with bounded high-output scrollback, prove sustained output while all six client xterms are warm-mounted but hidden causes zero renderer scheduler work, then drive 180 timed real-PTY spinner frames through a second production `session.tabs.subscribeAll` transport and the real Zustand store. Each worktree must publish and apply only its exact initial-working and final-idle titles. Both topologies ordinary-park five xterms without enabling the lossy retention budget, assert bounded cells/heap/timer lag, then restore the bounded authoritative tail exactly once on the original PTY including output produced while parked and continued input/output. The headed oracle additionally keeps every remote target workspace and terminal unmounted in the host renderer before and after client recovery. Production-parity React contracts reproduce and remove the independent cold-park watcher/title feedback loop, including memo-cache discard, StrictMode cleanup replay, detached split-PTY replacement, and the zero-park fast path. A deterministic cross-boundary contract drives 24 real PTY trackers in distinct worktrees through 2,880 decorative spinner and bare-native redraw chunks and counts host publications, serialized bytes, renderer apply calls, and store mutations while preserving exact per-PTY raw chunks and per-worktree semantic transitions. Fake-clock continuation beyond the 30-minute stale boundary requires one deduplicated refresh per worktree every 15 minutes, globally spaced by at least 50 ms, with every mirrored working status still fresh. Exact-identical Pi plus rotating normalized Pi/Grok frames renew coarse working/permission state; rich prompt/tool/question fields retain their original 30-minute lease, and `stateStartedAt`/acknowledgement identity remain stable across title/hook source swaps. A separate headed desktop-client journey relaunches the host main process while retaining the daemon and a daemon-backgrounded PTY, then requires first-observation foreground sync and continued input/output on the same PTY. Unit coverage keeps layout versions out of snapshot output high-waters when a restored snapshot has no output sequence and retries an existing subscriber only after authoritative provider inventory proves the retained PTY exists. Separate headed/headless ACK-starvation tests recover one stalled stream without replacing its PTY. A headed output-drop oracle proves a successful snapshot probe cannot certify a stale live stream when the authoritative PTY sequence advanced beyond the client high-water. Unit coverage requires an unsequenced client to establish a first-probe baseline, remain attached at the same sequence, and recover only after a later probe advances. A capability-disabled run proves legacy hosts retain the prior lossy limit/TTL fallback. Unit tests cover exact-owner capability routing, raw-stream release, singleton side-effect facts with timed handoff cleanup, 64 simultaneous synthetic-title sources with zero decorative client events after convergence, per-client late-subscribe recovery, semantic bell/title transitions, local animation-frame preservation, legacy no-output-pause delivery gating with exact snapshot restore, provider-authoritative snapshots, 128 active streams plus retry after a 129th-stream capacity rejection, capacity-pressure backoff, full split-leaf remint reconciliation, truncation, and manual-disconnect queue fencing. Linux, Windows, live SSH, and production-scale paired hosts remain gaps. The audit removed the E2E parking override/debug-handle suite. Override exposure guards, delay/retention independence, TTL filtering and debug-handle forwarding alarms are lost. Retained real parking journeys do not establish equivalent test-harness configuration coverage. The test-value audit removes separate responsive-stream, intervening-control-frame and capacity-error classification examples. Retained multiplexer checks use fake clocks and captured transport callbacks; their per-stream isolation, deadline and probe-baseline assertions do not prove an actual host restart or equivalent replacement coverage. Historical run commands and results remain historical. The test-value audit removes the standalone retained/live metadata projection and additive-preference negative example from runtime-client-events.test.ts. Its remaining case checks partial-authority event admission through a captured callback; it does not establish full-pair projection, actual peer delivery or equivalent replacement coverage. Historical commands and results remain historical. The isolated accepted-handle-events observer suite is removed. Its runtime/worktree/pane membership, once-only delivery and listener-release ledger alarms are lost. Live hidden-retention evidence does not establish equivalent observer coverage. Audit1292: The entire host-session mirror empty-inventory settle suite is removed. Live-unpublished versus genuinely-empty hydration, same-ID revision response/in-flight fencing and cleared-tracking pending-probe invalidation alarms are lost. The singleton store and inventories were mocked; no real host census or resume delivery or equivalent replacement coverage is claimed. Audit1302: src/main/runtime/session-tabs-inventory-publication.test.ts: Keep actual runtime public inventory authority across renderer epochs/resync, fallback, aborts and invalid graph publishers. Remove copied nonempty publication, private scan counters/collector and known-host replacement, fake refresh that performs the restoration itself, environment capability flag, and private resume-store characterization whose fixture supplies reservation/rollback ordering. Lose isolated refresh/churn/receiver/capability and copied-payload wiring alarms; no equivalent replacement. Reduced selectors require fresh validation; prior full-suite receipts do not validate removed or changed coverage. No equivalent replacement or physical-provider claim. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1318: src/shared/remote-runtime/control/remote-runtime-shared-control-connection.test.ts: Retain real local encrypted WebSocket regressions for pre-ready socket replacement preserving a request, close-before-ready logical subscription cleanup, unrelated RPC surviving another deadline, and pending setup abort leaving a sibling pending. Host server is synthetic and uses production crypto; kept scenarios also retain internal diagnostics/memory assertions unchanged. Remove happy routing, authority-envelope, unknown-ID/log, cleanup enum, binary, keepalive/liveness and production-budget/private socket inventories; intentionally lose those alarms, no physical host authorization, independent crypto or remote provider claim. Reduced selectors need fresh validation; historical full-suite receipts do not validate the changed coverage. No equivalent replacement or physical-provider claim. Audit1346 removes the entire private-state-injected session-tabs rich-status boundary suite. Its stale task/question, neutral-title preservation, owner-change and provider-generation alarms are intentionally lost. Retained real terminal journeys do not establish equivalent replacement coverage. The prior mixed 42-contract execution receipt is historical and must not validate the reduced current selector; fresh validation remains required. Audit1352: focused Linux unit selections validated on source head b638e07d395b76af3ba6b03761a1a0fe7e5e026e. Historical statements above retain their original scope. Intentional test losses are recorded in knownGaps; these runs claim no physical-provider, p95, rendered E2E or cross-platform validation. Audit1360: focused Linux unit selections validated on source head 415c9cffc2d10024131ec4b6721767a22195de6d. Historical statements above retain their original scope. Intentional test losses are recorded in knownGaps; these runs claim no physical-provider, p95, rendered E2E or cross-platform validation. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1391: src/renderer/src/components/terminal-pane/pty-connection-remote-pause-drop-notice.spec.ts reduced. Lose repeated quiet switches avoiding snapshots and legacy host without drop-report capability requesting a snapshot. Retained imported registrar uses fake transport/xterm writes, exercising real gap/restore lifecycle rather than actual terminal rendering. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1396: src/renderer/src/components/terminal-pane/terminal-parked-watcher-reconciliation.test.ts reduced. Retire standalone reconcile set output and positive untouchedFreshSpawn/workspaceOwner copying alarms; no native watcher execution claim. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1402: src/renderer/src/components/terminal-pane/pty-connection-park-reveal-model-snapshot.spec.ts reduced. Loses basic paired serialize/initial replay forwarding check; hidden alt-frame grid/precedence policy remains. No real terminal renderer claim. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1412: src/main/runtime/mobile-session-tabs-agent-status-heartbeat.test.ts reduced. Loses FIFO fairness/spacing, semantic bypass/cancellation, delayed-boundary scheduling and teardown timer-cleanup alarms. Retained fixed emission cases do not replace those outcomes. No equivalent replacement coverage is claimed.", "motivatingLinks": [ "https://github.com/stablyai/orca/issues/8652", "https://github.com/stablyai/orca/pull/10625" ], "invariant": "A host advertising terminal.paired-parking.v1 keeps the PTY and bounded authoritative history alive while an ordinary hidden-view park destroys the client xterm and releases its raw per-PTY stream. Reveal must restore up to the requested 5,000 rows, parked-time side effects/output, the same PTY identity, and continued input/output. Parked watcher synchronization compares parking semantics rather than fresh object identity, owns setup and cleanup across StrictMode replay, reconciles real PTY/layout changes, and performs no terminal-state scan when nothing is parked. After a host main-process relaunch, unknown local delivery-sync state must not be treated as proof that a surviving daemon PTY is already foregrounded. An empty headed session-tab inventory is authoritative only after the current renderer graph generation publishes a complete inventory and every execution host answers the PTY census; unpublished, resync-pending, transport-failed, and partially scoped inventories remain unverifiable, while genuinely authoritative empty and non-empty inventories settle and a scope-less legacy empty keeps its pre-authority best-effort settle so outdated hosts retain agent auto-resume. The unpublished-empty regression must keep the resume dispatch parked with exactly one daemon process/writer; restoring unconditional empty hydration must release one dispatch and create two live writers. If subscription precedes provider readiness, authoritative inventory must reconsider that existing subscriber without spawning, resizing, or changing reconnect state. A snapshot without an output high-water must keep it unknown rather than borrowing a layout version that can suppress newer live output. Hosts without the capability must gate hidden raw output before xterm scheduling and repaint from the authoritative snapshot on reveal while retaining the existing limit/TTL force-parking fallback. After the first semantic title state, decorative spinner frequency must add zero paired client-event frames and zero full session-tabs work between status-freshness leases regardless of hidden worktree count. Continuous working or permission evidence may refresh each affected worktree once per 15 minutes through one globally spaced FIFO so current and legacy viewers do not decay before 30 minutes; sibling PTYs share that worktree refresh, while local title animation, raw output, and semantic title, status, bell, completion, and query facts remain intact. A paired terminal stream whose delivery credits stop progressing must replace only that stream; command silence first probes authoritative state and replaces the stream if the probe times out or proves the PTY advanced beyond the client's delivered output sequence. When no comparable delivery high-water exists, the first sequenced probe establishes a baseline and only later advancement proves staleness. A same-sequence snapshot remains valid proof of a responsive silent command. A successful status probe may replace a pre-ready shared-control socket without rejecting or duplicating calls already waiting for that transport. Manual disconnect must retain pairing while preventing queued or passive calls and subscriptions from recreating transport until explicit Connect. A paired pane reconnecting to a relaunched host that has not yet published its window graph (an unpublished frame, bare or client-projected) must stay in recovery rather than retire, and the host's first published snapshot carrying the surface must reattach it; a published frame lacking the surface still retires it. The placeholder is only ever sent before the host's graph first publishes; afterwards a worktree with no tabs gets a real empty answer, pushed once to any client that asked during startup, so a paired client opening a never-opened host worktree still gets its first terminal.", - "oracle": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. The direct side-effect-facts selector retains six cases: replay exclusion, live sequence handling, stale disposer ownership, one handoff drain, drain-before-snapshot, and stale replacement-snapshot refusal. Successful current-entry snapshot application is not witnessed by the retained snapshot case. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2003: Remove SSH/absent-inventory attach refusal, later-subscriber retry, no-view adoption/takeback, provider-tail/unknown-read projections, failed-keystroke retry and simpler loss reattach examples. Keep the existing-subscriber inventory-ready retry alias only at its retained callback, plus retained held-input cancellation/replacement and overlapping attach/loss fences. The inventory driver also loses the whole empty-worktree-publication registration and folder-host/unverifiable scalar examples. Retire the entire dated10-case attach receipt; retain16 independent receipts only in their original scopes. No retained-memory or real-daemon measurement follows from supplied provider inventories. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1991: Remove initial-snapshot sequence-domain, limited-tail/truncation and unknown-output-sequence aliases whose callbacks are removed. The source retains partial-chunk replay and requested-snapshot retry/admission at supplied snapshot/data seams, not equivalent replacement or measured retained-memory proof. Preserve the existing stream/pending-wait capacity aliases only to their retained owning multiplex callbacks with supplied PTY resolutions. Removing the sibling-PTY heartbeat dedup work-count sample does not establish a new decorative fanout budget; the unrelated-PTY cancellation fence survives. Preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Remove the reasonless-close selector and narrow census aliases to actual retained runtime and injected-inventory callbacks. The removed synthetic census-dedup and isolated pre-boundary timing cases receive no replacement credit; preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1939 (sources1934/1935): Delete the synthetic watcher-synchronization and partial-reconciliation suites. Retire all affected host-fact, capability-owner, activation-residue and retention-cap aliases because the remaining callbacks do not establish those full claims. Keep exact local hook/timer, byte-stream side effects, exit/disposal, scroll-intent cleanup and deterministic local rank/TTL progression callbacks within their mocked-host scope. No physical paired-host retention, resource budget or equivalent replacement evidence is assigned. The generic broad test command remains a runnable discovery command, never a dated execution receipt. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete hook-agent-status projection suite and its two behavioral assertion aliases. Supplied stored hook/title/resume/staleness/rename/question/generation/PTY-retirement/snapshot-version projections are intentionally lost. The separate osc-status-scan gate mentions this source only in existing historical notes and receives no mutation. No measured hidden-output retention, CPU/RSS or resource budget result is claimed by the deleted source or this audit. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Retained exact unchanged-source receipts remain the only current execution evidence. Reduced selectors have no fresh execution result. Historical pre-audit oracle account for changed sources: The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Run one byte-identical six-terminal oracle against an isolated headed desktop host and an isolated headless `orca serve` host. Stage at least 1,000,000 xterm cells, hide all six warm managers, emit an initial working title followed by 30 timed real-PTY spinner frames and a final idle title per worktree, and observe a second real session-tabs subscription plus the production renderer store. Require exactly two transport envelopes and two title mutations per worktree, no decorative title crossing either boundary, and positive decoded-envelope bytes. Enable ordinary parking with the lossy retention budget disabled, require exactly one mounted manager and five parked tabs, at most 45% retained cells, no more than 16 MiB heap growth, and under 500 ms timer drift. In headed mode, require the host renderer to remain on its original workspace with zero target terminal managers mounted throughout client park, reveal, and live I/O. While a tab is parked, require authoritative terminal.read to observe new PTY output; reveal it and require the original PTY, a marker within the requested 5,000-row history, the parked marker, and post-reveal input/output. Drive 24 real PTY title trackers in distinct worktrees through 40 decorative spinner frames each, then 40 spinner frames alternating with Cursor's bare native identity redraw, using fake clocks. Require all 2,880 raw chunks exactly and zero host session-tabs publications, serialized bytes, renderer apply calls, and renderer store mutations. Continue decorative evidence beyond 30 minutes, require exactly one refresh per worktree per 15-minute lease through a FIFO spaced by at least 50 ms, and require all 24 viewer statuses to remain working and fresh; then require one exact semantic idle transition and one visible-output chunk per PTY. Require ordinary and interior-Braille title changes to publish. Drive 64 host PTYs through ten 80 ms synthetic title frames, require zero paired client events after the first frame per PTY while all 640 local frames remain observable, attach a late client and require one current frame per PTY, then require semantic bell and idle transitions on both clients. Reconstruct a legacy subscribed stream with no outputPause capability, hide a chatty pane, require no hidden xterm writes, reveal it, and require one authoritative snapshot plus continued live output. Drop and acknowledge output only for one original paired-client stream, prove the fixture process consumed input while the host model advanced and the client stayed stale, then require the command snapshot probe to replace that stream, repaint exact fixture output, preserve authoritative PTY identity and target-tab cardinality, and resume live I/O. For a client without a delivered sequence, require the first numeric probe to establish a baseline without replacement, a same-sequence probe to remain attached, and a later advanced probe to replace only that stream. Withhold the first encrypted shared-control ready frame, start one RPC, trigger a status-probe refresh, and require two connections, one host delivery, successful response, and zero retained request bytes. Admit 128 active streams, reject the 129th as retryable, release one stream, then require the retry to attach and publish a snapshot without multiplying retained subscribers. Disable terminal.paired-parking.v1 and require the same oracle to fail before parking, while the legacy limit-one fallback separately passes. Also preserve truncated first paint, ACK-starved same-PTY recovery, responsive silent-command snapshot probes, dead-stream probe timeout recovery, and queued manual-disconnect fencing.", + "oracle": "Audit 3318 reduces remote-runtime-pty-transport-unpublished-host-graph.test.ts, retiring its direct published-empty-graph omission-to-pane-exit alarm. The unchanged bounded unpublished-frame/pushed-recovery and stale-handle park/manual-retry cases are narrower, not equivalent replacements. Both complete three-case Linux receipts dated 2026-11-04 are archived as historical, including source heads; they do not validate this reduced selector. Remaining fourteen active receipts are macOS only. No fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. The direct side-effect-facts selector retains six cases: replay exclusion, live sequence handling, stale disposer ownership, one handoff drain, drain-before-snapshot, and stale replacement-snapshot refusal. Successful current-entry snapshot application is not witnessed by the retained snapshot case. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2003: Remove SSH/absent-inventory attach refusal, later-subscriber retry, no-view adoption/takeback, provider-tail/unknown-read projections, failed-keystroke retry and simpler loss reattach examples. Keep the existing-subscriber inventory-ready retry alias only at its retained callback, plus retained held-input cancellation/replacement and overlapping attach/loss fences. The inventory driver also loses the whole empty-worktree-publication registration and folder-host/unverifiable scalar examples. Retire the entire dated10-case attach receipt; retain16 independent receipts only in their original scopes. No retained-memory or real-daemon measurement follows from supplied provider inventories. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1991: Remove initial-snapshot sequence-domain, limited-tail/truncation and unknown-output-sequence aliases whose callbacks are removed. The source retains partial-chunk replay and requested-snapshot retry/admission at supplied snapshot/data seams, not equivalent replacement or measured retained-memory proof. Preserve the existing stream/pending-wait capacity aliases only to their retained owning multiplex callbacks with supplied PTY resolutions. Removing the sibling-PTY heartbeat dedup work-count sample does not establish a new decorative fanout budget; the unrelated-PTY cancellation fence survives. Preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, physical provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Remove the reasonless-close selector and narrow census aliases to actual retained runtime and injected-inventory callbacks. The removed synthetic census-dedup and isolated pre-boundary timing cases receive no replacement credit; preserve17 unrelated dated receipts. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1939 (sources1934/1935): Delete the synthetic watcher-synchronization and partial-reconciliation suites. Retire all affected host-fact, capability-owner, activation-residue and retention-cap aliases because the remaining callbacks do not establish those full claims. Keep exact local hook/timer, byte-stream side effects, exit/disposal, scroll-intent cleanup and deterministic local rank/TTL progression callbacks within their mocked-host scope. No physical paired-host retention, resource budget or equivalent replacement evidence is assigned. The generic broad test command remains a runnable discovery command, never a dated execution receipt. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete hook-agent-status projection suite and its two behavioral assertion aliases. Supplied stored hook/title/resume/staleness/rename/question/generation/PTY-retirement/snapshot-version projections are intentionally lost. The separate osc-status-scan gate mentions this source only in existing historical notes and receives no mutation. No measured hidden-output retention, CPU/RSS or resource budget result is claimed by the deleted source or this audit. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Retained exact unchanged-source receipts remain the only current execution evidence. Reduced selectors have no fresh execution result. Historical pre-audit oracle account for changed sources: The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Run one byte-identical six-terminal oracle against an isolated headed desktop host and an isolated headless `orca serve` host. Stage at least 1,000,000 xterm cells, hide all six warm managers, emit an initial working title followed by 30 timed real-PTY spinner frames and a final idle title per worktree, and observe a second real session-tabs subscription plus the production renderer store. Require exactly two transport envelopes and two title mutations per worktree, no decorative title crossing either boundary, and positive decoded-envelope bytes. Enable ordinary parking with the lossy retention budget disabled, require exactly one mounted manager and five parked tabs, at most 45% retained cells, no more than 16 MiB heap growth, and under 500 ms timer drift. In headed mode, require the host renderer to remain on its original workspace with zero target terminal managers mounted throughout client park, reveal, and live I/O. While a tab is parked, require authoritative terminal.read to observe new PTY output; reveal it and require the original PTY, a marker within the requested 5,000-row history, the parked marker, and post-reveal input/output. Drive 24 real PTY title trackers in distinct worktrees through 40 decorative spinner frames each, then 40 spinner frames alternating with Cursor's bare native identity redraw, using fake clocks. Require all 2,880 raw chunks exactly and zero host session-tabs publications, serialized bytes, renderer apply calls, and renderer store mutations. Continue decorative evidence beyond 30 minutes, require exactly one refresh per worktree per 15-minute lease through a FIFO spaced by at least 50 ms, and require all 24 viewer statuses to remain working and fresh; then require one exact semantic idle transition and one visible-output chunk per PTY. Require ordinary and interior-Braille title changes to publish. Drive 64 host PTYs through ten 80 ms synthetic title frames, require zero paired client events after the first frame per PTY while all 640 local frames remain observable, attach a late client and require one current frame per PTY, then require semantic bell and idle transitions on both clients. Reconstruct a legacy subscribed stream with no outputPause capability, hide a chatty pane, require no hidden xterm writes, reveal it, and require one authoritative snapshot plus continued live output. Drop and acknowledge output only for one original paired-client stream, prove the fixture process consumed input while the host model advanced and the client stayed stale, then require the command snapshot probe to replace that stream, repaint exact fixture output, preserve authoritative PTY identity and target-tab cardinality, and resume live I/O. For a client without a delivered sequence, require the first numeric probe to establish a baseline without replacement, a same-sequence probe to remain attached, and a later advanced probe to replace only that stream. Withhold the first encrypted shared-control ready frame, start one RPC, trigger a status-probe refresh, and require two connections, one host delivery, successful response, and zero retained request bytes. Admit 128 active streams, reject the 129th as retryable, release one stream, then require the retry to attach and publish a snapshot without multiplying retained subscribers. Disable terminal.paired-parking.v1 and require the same oracle to fail before parking, while the legacy limit-one fallback separately passes. Also preserve truncated first paint, ACK-starved same-PTY recovery, responsive silent-command snapshot probes, dead-stream probe timeout recovery, and queued manual-disconnect fencing.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-unpublished-host-graph.test.ts", "ORCA_BACKGROUND_LAUNCH=1 pnpm exec playwright test tests/e2e/paired-remote-terminal-host-quit-reconnect-input.spec.ts --config tests/playwright.config.ts --project electron-headless --workers=1", @@ -8871,26 +8872,6 @@ "durationSeconds": 17.5, "summary": "The headed six-PTY oracle kept all warm-mounted hidden streams at zero renderer scheduler enqueues/drains under sustained output and under 500 ms timer lag, then restored the authoritative flood tail and parked/live markers exactly once.", "currentCommand": "ORCA_E2E_WEB_CLIENT=1 SKIP_BUILD=1 pnpm exec playwright test tests/e2e/paired-remote-terminal-retention-memory.spec.ts --config tests/playwright.config.ts --project electron-headful --workers=1" - }, - { - "date": "2026-11-04", - "runner": "local", - "platform": "linux", - "result": "passed", - "command": "ORCA_BACKGROUND_LAUNCH=1 pnpm --config.verifyDepsBeforeRun=false test src/renderer/src/components/terminal-pane/remote-runtime-pty-transport-unpublished-host-graph.test.ts --maxWorkers=2", - "currentCommand": "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-unpublished-host-graph.test.ts", - "durationSeconds": 3.87, - "summary": "Test Files 1 passed (1); Tests 3 passed (3). Focused Linux unit execution; no physical-provider, p95 or cross-platform claim. Source head b638e07d395b76af3ba6b03761a1a0fe7e5e026e." - }, - { - "date": "2026-11-04", - "runner": "local", - "platform": "linux", - "result": "passed", - "command": "ORCA_BACKGROUND_LAUNCH=1 pnpm --config.verifyDepsBeforeRun=false test src/renderer/src/components/terminal-pane/remote-runtime-pty-transport-unpublished-host-graph.test.ts --maxWorkers=2", - "currentCommand": "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-unpublished-host-graph.test.ts", - "durationSeconds": 3.79, - "summary": "Test Files 1 passed (1); Tests 3 passed (3). Focused Linux unit execution; no physical-provider, p95 or cross-platform claim. Source head 415c9cffc2d10024131ec4b6721767a22195de6d." } ], "runtimeBudget": { @@ -9019,7 +9000,10 @@ "Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical.", "Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed.", "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed.", - "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed." + "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed.", + "Audit 3318 reduces remote-runtime-pty-transport-unpublished-host-graph.test.ts, retiring its direct published-empty-graph omission-to-pane-exit alarm. The unchanged bounded unpublished-frame/pushed-recovery and stale-handle park/manual-retry cases are narrower, not equivalent replacements. Both complete three-case Linux receipts dated 2026-11-04 are archived as historical, including source heads; they do not validate this reduced selector. Remaining fourteen active receipts are macOS only. No fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed.", + "Audit 3318 retires this complete dated three-case Linux unit receipt because remote-runtime-pty-transport-unpublished-host-graph.test.ts was reduced. Original command/currentCommand/result/date/duration/source head/count remain historical and do not validate the reduced selector. Original receipt verbatim:\n{\n \"date\": \"2026-11-04\",\n \"runner\": \"local\",\n \"platform\": \"linux\",\n \"result\": \"passed\",\n \"command\": \"ORCA_BACKGROUND_LAUNCH=1 pnpm --config.verifyDepsBeforeRun=false test src/renderer/src/components/terminal-pane/remote-runtime-pty-transport-unpublished-host-graph.test.ts --maxWorkers=2\",\n \"currentCommand\": \"pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-unpublished-host-graph.test.ts\",\n \"durationSeconds\": 3.87,\n \"summary\": \"Test Files 1 passed (1); Tests 3 passed (3). Focused Linux unit execution; no physical-provider, p95 or cross-platform claim. Source head b638e07d395b76af3ba6b03761a1a0fe7e5e026e.\"\n }", + "Audit 3318 retires this complete dated three-case Linux unit receipt because remote-runtime-pty-transport-unpublished-host-graph.test.ts was reduced. Original command/currentCommand/result/date/duration/source head/count remain historical and do not validate the reduced selector. Original receipt verbatim:\n{\n \"date\": \"2026-11-04\",\n \"runner\": \"local\",\n \"platform\": \"linux\",\n \"result\": \"passed\",\n \"command\": \"ORCA_BACKGROUND_LAUNCH=1 pnpm --config.verifyDepsBeforeRun=false test src/renderer/src/components/terminal-pane/remote-runtime-pty-transport-unpublished-host-graph.test.ts --maxWorkers=2\",\n \"currentCommand\": \"pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-unpublished-host-graph.test.ts\",\n \"durationSeconds\": 3.79,\n \"summary\": \"Test Files 1 passed (1); Tests 3 passed (3). Focused Linux unit execution; no physical-provider, p95 or cross-platform claim. Source head 415c9cffc2d10024131ec4b6721767a22195de6d.\"\n }" ], "demotionRule": "Keep experimental or demote to protection none if decorative title frequency changes paired client-event or session-tabs publication volume after convergence, a semantic fact is dropped, paired first paint is blank, passive work reconnects after manual disconnect, a capable hidden paired manager survives ordinary parking, a legacy hidden pane schedules raw xterm output or bypasses its fallback, reveal loses requested bounded history or live PTY identity, retained buffer cells do not fall, or either headed/headless scenario flakes." }, @@ -9198,7 +9182,7 @@ "providers": ["local", "daemon", "ssh", "remote-runtime"], "coveredPlatforms": ["macos"], "coveredProviders": [], - "coverageNotes": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Module layout SSH policy: Retire the multiplexer private-clock saturation registration, the writer saturation transition-count registration and the changed-source dated writer/backpressure receipt. Retained byte delivery, lane order, fairness, settlement and cancellation cases are unvalidated by that historical receipt. No equivalent replacement, new execution, measured performance, live SSH/WSL or cross-platform proof is claimed. The unchanged WSL protocol receipt retains only its original command scope. Historical prior account (preserved verbatim): Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2070 (sources2068/2069): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/hooks-and-hosted-review.spec'). src/main/ssh/ssh-multiplexer-transport-writer.test.ts: reports saturated epochs and their drain boundary exactly once. src/main/runtime/orca-runtime-tests/worktree-removal-archive-hook-gate.spec.ts: removes and records the waiver when the failure is explicitly overridden; removes without an override record when the hook succeeds; removes when the hook is configured but not requested; removes when no archive hook is configured; holds a desktop removal in the same repo until this one is accepted; joins a desktop removal of the same worktree accepted while this one waited. src/main/runtime/orca-runtime-tests/worktree-ps-agent-row-dismissal.spec.ts: drops the row as soon as the user dismisses it, without waiting for the PTY to exit; tells paired clients to republish on the transition and on the dismissal; rejoins a row through its %s handle after pane ownership clears. src/main/runtime/orca-runtime-tests/hooks-and-hosted-review.spec.ts: reports an error when the SSH filesystem provider is unavailable; reports an SSH orca.yaml as an error without reading any filesystem; reports ok for a local repo hook check; reads SSH issue commands as absent and drops writes without touching any filesystem; allows host integration slug helpers for SSH repos through provider-aware GitHub clients; routes runtime GitHub repo identity helpers through the selected WSL project runtime; routes runtime GitHub issue and work-item actions through the selected WSL project runtime; pins explicit origin preference on runtime open-by-number work item lookups; forwards check-details cancellation without local Git overrides. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. The historical receipt/currentCommand selecting src/main/ssh/ssh-multiplexer-transport-writer.test.ts remains historical and is not a passing result for the reduced current command. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete tail-wait memo suite: its sole current authored callback checks supplied fromTail/blocked field values. Retire all three broader memoization/equivalence/eviction assertion aliases; they are not backed by retained source callbacks. No scan-work budget, sustained output, physical terminal or resource measurement is assigned to this removed test. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1674 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Local macOS evidence covers main-process pending-output and projection-admission caps, saturated-tail wait detection, direct-SSH source retention/accounting, always-on V1 negotiation with legacy peer fallback, exact provider-generation pause ownership, renderer exit/data ordering, decoder input bounds, bounded reconnect scheduling, and WSL stdio transport settlement through deterministic unit contracts. A joined main/runtime oracle exercises renderer-sourced headed semantics and headless-model snapshot semantics without launching a live paired server. Separately, a macOS-hosted Docker OpenSSH run exercises only the deployed Linux relay and direct SSH provider with an exact 256 KiB source-credit plateau, concurrent PTY typing, fixed-size filesystem frames, Git churn, and owner-lease reconnect. Neither deterministic topology labels nor Docker SSH constitute live headed desktop, headless orca serve, or physical WSL evidence; local/daemon, prior-version daemons, Windows named pipes/ConPTY, folder workspaces, and mixed-version clients also remain uncovered. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result.", + "coverageNotes": "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Module layout SSH policy: Retire the multiplexer private-clock saturation registration, the writer saturation transition-count registration and the changed-source dated writer/backpressure receipt. Retained byte delivery, lane order, fairness, settlement and cancellation cases are unvalidated by that historical receipt. No equivalent replacement, new execution, measured performance, live SSH/WSL or cross-platform proof is claimed. The unchanged WSL protocol receipt retains only its original command scope. Historical prior account (preserved verbatim): Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2070 (sources2068/2069): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/hooks-and-hosted-review.spec'). src/main/ssh/ssh-multiplexer-transport-writer.test.ts: reports saturated epochs and their drain boundary exactly once. src/main/runtime/orca-runtime-tests/worktree-removal-archive-hook-gate.spec.ts: removes and records the waiver when the failure is explicitly overridden; removes without an override record when the hook succeeds; removes when the hook is configured but not requested; removes when no archive hook is configured; holds a desktop removal in the same repo until this one is accepted; joins a desktop removal of the same worktree accepted while this one waited. src/main/runtime/orca-runtime-tests/worktree-ps-agent-row-dismissal.spec.ts: drops the row as soon as the user dismisses it, without waiting for the PTY to exit; tells paired clients to republish on the transition and on the dismissal; rejoins a row through its %s handle after pane ownership clears. src/main/runtime/orca-runtime-tests/hooks-and-hosted-review.spec.ts: reports an error when the SSH filesystem provider is unavailable; reports an SSH orca.yaml as an error without reading any filesystem; reports ok for a local repo hook check; reads SSH issue commands as absent and drops writes without touching any filesystem; allows host integration slug helpers for SSH repos through provider-aware GitHub clients; routes runtime GitHub repo identity helpers through the selected WSL project runtime; routes runtime GitHub issue and work-item actions through the selected WSL project runtime; pins explicit origin preference on runtime open-by-number work item lookups; forwards check-details cancellation without local Git overrides. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. The historical receipt/currentCommand selecting src/main/ssh/ssh-multiplexer-transport-writer.test.ts remains historical and is not a passing result for the reduced current command. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete tail-wait memo suite: its sole current authored callback checks supplied fromTail/blocked field values. Retire all three broader memoization/equivalence/eviction assertion aliases; they are not backed by retained source callbacks. No scan-work budget, sustained output, physical terminal or resource measurement is assigned to this removed test. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1674 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Local macOS evidence covers main-process pending-output and projection-admission caps, saturated-tail wait detection, direct-SSH source retention/accounting, always-on V1 negotiation with legacy peer fallback, exact provider-generation pause ownership, renderer exit/data ordering, decoder input bounds, bounded reconnect scheduling, and WSL stdio transport settlement through deterministic unit contracts. A joined main/runtime oracle exercises renderer-sourced headed semantics and headless-model snapshot semantics without launching a live paired server. Separately, a macOS-hosted Docker OpenSSH run exercises only the deployed Linux relay and direct SSH provider with an exact 256 KiB source-credit plateau, concurrent PTY typing, fixed-size filesystem frames, Git churn, and owner-lease reconnect. Neither deterministic topology labels nor Docker SSH constitute live headed desktop, headless orca serve, or physical WSL evidence; local/daemon, prior-version daemons, Windows named pipes/ConPTY, folder workspaces, and mixed-version clients also remain uncovered. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result.", "motivatingLinks": [ "https://github.com/stablyai/orca/pull/6836", "https://github.com/stablyai/orca/pull/6858", @@ -9206,7 +9190,7 @@ "https://github.com/stablyai/orca/pull/7054" ], "invariant": "High-volume terminal output must stay bounded across daemon socket writes, SSH relay writer/decoder/source retention/reconnect, main model and projection admission, runtime metadata, detectors, and tail checks, main-to-renderer in-flight bytes, renderer scheduler queues, and hidden-output restore without starving focused input. Every new SSH session offers V1 and every same-build relay supports it; legacy delivery is reachable only when capability negotiation proves the peer cannot use V1. Negotiated SSH spans remain exact and contiguous through activation, exit, cancellation, recovery, desktop projection, and required remote replacement; provisional activation data cannot project, reconnect may release it only into the exact attempt's private recovery quarantine after contract validation, and exit seals that quarantine against later same-token frames until ordered admission or exact cancellation proof. Reconnect checkpoint capture freezes exact provider-generation/PTy admission, cancels queued old work, and waits for the running raw completion before exposing its accepted checkpoint; timeout or failure detaches only that PTy's old model, makes the checkpoint unavailable, and releases the bounded fence without closing the shared provider, while the same failure outside an active exact migration remains generation-fatal and overlapping reconnects preserve any earlier outstanding fence. Pending renderer projection IDs are capped and compacted across split remainders, exit preparation owns its renderer fence through finalization, exit-time cancellation transfers published projections before proof commit, generation close fences late proof while draining exact projection waiters, closed-generation identity compacts without weakening stale rejection, canceled source-delivery retirement retains at most one ordered token per PTY, and pause/resume targets only the exact provider generation. Additional subscribers cannot stall the owner, while the required legacy primary retains backpressure.", - "oracle": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Module layout SSH policy: Retire the multiplexer private-clock saturation registration, the writer saturation transition-count registration and the changed-source dated writer/backpressure receipt. Retained byte delivery, lane order, fairness, settlement and cancellation cases are unvalidated by that historical receipt. No equivalent replacement, new execution, measured performance, live SSH/WSL or cross-platform proof is claimed. The unchanged WSL protocol receipt retains only its original command scope. Historical prior account (preserved verbatim): Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2070 (sources2068/2069): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/hooks-and-hosted-review.spec'). src/main/ssh/ssh-multiplexer-transport-writer.test.ts: reports saturated epochs and their drain boundary exactly once. src/main/runtime/orca-runtime-tests/worktree-removal-archive-hook-gate.spec.ts: removes and records the waiver when the failure is explicitly overridden; removes without an override record when the hook succeeds; removes when the hook is configured but not requested; removes when no archive hook is configured; holds a desktop removal in the same repo until this one is accepted; joins a desktop removal of the same worktree accepted while this one waited. src/main/runtime/orca-runtime-tests/worktree-ps-agent-row-dismissal.spec.ts: drops the row as soon as the user dismisses it, without waiting for the PTY to exit; tells paired clients to republish on the transition and on the dismissal; rejoins a row through its %s handle after pane ownership clears. src/main/runtime/orca-runtime-tests/hooks-and-hosted-review.spec.ts: reports an error when the SSH filesystem provider is unavailable; reports an SSH orca.yaml as an error without reading any filesystem; reports ok for a local repo hook check; reads SSH issue commands as absent and drops writes without touching any filesystem; allows host integration slug helpers for SSH repos through provider-aware GitHub clients; routes runtime GitHub repo identity helpers through the selected WSL project runtime; routes runtime GitHub issue and work-item actions through the selected WSL project runtime; pins explicit origin preference on runtime open-by-number work item lookups; forwards check-details cancellation without local Git overrides. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. The historical receipt/currentCommand selecting src/main/ssh/ssh-multiplexer-transport-writer.test.ts remains historical and is not a passing result for the reduced current command. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete tail-wait memo suite: its sole current authored callback checks supplied fromTail/blocked field values. Retire all three broader memoization/equivalence/eviction assertion aliases; they are not backed by retained source callbacks. No scan-work budget, sustained output, physical terminal or resource measurement is assigned to this removed test. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1674 retires changed-source assertion and execution scope without equivalent replacement coverage. Existing unchanged-selector receipts remain scoped to their original sources. Historical prior-source oracle: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: The current executable slice injects main-process renderer backlog pressure, then asserts ordinary terminal-wait detection without a joined-tail allocation and with full prompt-family continuity, per-PTY and total pending-output caps, a 1,024-ID projection cap with split-remainder compaction, preserved sequenced-tail metadata, active-pending protection ahead of background trimming, exact provider-generation pause ownership, renderer exit/data ordering, and ACK-gated in-flight bounds. The negotiation seam proves every initial connection and reconnect offers V1, the same-build relay advertises V1 without launch flags, old clients remain token-free, and method-not-found peers fall back without installing source ACK publication. The joined source-intake/multiplex oracle blocks renderer or headless serialization, admits one snapshot-covered span and one trailing span, requires replacement reservation only after the authoritative sequence is known, advances upstream ACK eligibility exactly to the covered boundary after SnapshotEnd, and delivers then ACKs the trailing span through the ordinary live path without cancellation. Source-range ledger and multiplex seams additionally require cumulative byte ACK 40 then 100 for one 100-byte frame to release exact byte credit without early source settlement, admit a contiguous higher-generation recovered token while the prior token remains unsettled, and reject stale stream or source generations without releasing in-flight byte credit. Replacement reservation rolls back every earlier span on a later failure, commit requires the exact frozen transfer states, and an authoritatively reclaimed covered span rejects commit without local trim authority while rollback removes the reservation idempotently. The exit-deadline oracle publishes a source-backed desktop projection, advances a fake clock to cancellation, and requires one projection transfer, proof commit, preparation, and final exit with no provider close, retained obligation, duplicate finalization, or process-lifetime cancellation tombstone; a second controlled proof remains pending across generation close and must never publish final exit, while both paths release the renderer preparation lease exactly once. The model-migration oracle accepts span A, blocks span B's raw emulator completion, begins disconnect migration, and proves no checkpoint or reconnect attach can remain at A; releasing B advances the exact checkpoint to B and projects B once. Its fake-clock arm reaches the 10-second deadline, requires checkpoint-unavailable, one exact model reset, zero retained admission charge and timers, and no effect from late raw settlement. Its two-PTY failure arm rejects one running raw callback under an active exact migration and requires one checkpoint-unavailable result, one model reset, zero provider closes, retained sibling checkpoint and transport liveness, zero charges and timers, and an unaffected other generation; a non-migrating sibling oracle still requires generation close. The intake additionally closes 2,048 sequential provider generations into one exact range, rejects stale events across that range, and preserves an out-of-order live gap until its own close. Direct-SSH contracts additionally prove activation-response settlement before the first source frame, claim-gated projection under same-decoder-turn response/data delivery, exact early-ACK reservation through send settlement or same-token retry, recovery completion visible before lease-held frames transfers those frames only into the private quarantine and admits no output until exact checkpoint-to-recoveryEnd coverage, invalid-checkpoint restore retention through response settlement followed by fresh-token retry and one live source frame, isolated saturated-subscriber eviction with healthy owner liveness, preserved legacy-primary backpressure, one-write ownership across write(false)/drain, control selection ahead of queued ordinary PTY input with FIFO lanes and bounded fairness, liveness rebasing, failed-exit retention through late ACK or exact recovery, response-settlement cancellation authority, token-scoped exit cancellation, exact checkpoint-to-recoveryEnd continuity, stale-owner one-shot retry, no physical PTY teardown on recovery failure, exact provider-generation closure and once-only cleanup when recovery cancellation publication or proof rejects, 16 MiB-plus-1 MiB decoder caps, charged per-PTY/session retention budgets, and eight-wide isolated reattach. The live Electron perf oracle adds hidden-output floods, renderer scheduler queue depth, dropped-output-zero normal scenarios, and active key latency budgets before promotion.", + "oracle": "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Module layout SSH policy: Retire the multiplexer private-clock saturation registration, the writer saturation transition-count registration and the changed-source dated writer/backpressure receipt. Retained byte delivery, lane order, fairness, settlement and cancellation cases are unvalidated by that historical receipt. No equivalent replacement, new execution, measured performance, live SSH/WSL or cross-platform proof is claimed. The unchanged WSL protocol receipt retains only its original command scope. Historical prior account (preserved verbatim): Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2070 (sources2068/2069): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/hooks-and-hosted-review.spec'). src/main/ssh/ssh-multiplexer-transport-writer.test.ts: reports saturated epochs and their drain boundary exactly once. src/main/runtime/orca-runtime-tests/worktree-removal-archive-hook-gate.spec.ts: removes and records the waiver when the failure is explicitly overridden; removes without an override record when the hook succeeds; removes when the hook is configured but not requested; removes when no archive hook is configured; holds a desktop removal in the same repo until this one is accepted; joins a desktop removal of the same worktree accepted while this one waited. src/main/runtime/orca-runtime-tests/worktree-ps-agent-row-dismissal.spec.ts: drops the row as soon as the user dismisses it, without waiting for the PTY to exit; tells paired clients to republish on the transition and on the dismissal; rejoins a row through its %s handle after pane ownership clears. src/main/runtime/orca-runtime-tests/hooks-and-hosted-review.spec.ts: reports an error when the SSH filesystem provider is unavailable; reports an SSH orca.yaml as an error without reading any filesystem; reports ok for a local repo hook check; reads SSH issue commands as absent and drops writes without touching any filesystem; allows host integration slug helpers for SSH repos through provider-aware GitHub clients; routes runtime GitHub repo identity helpers through the selected WSL project runtime; routes runtime GitHub issue and work-item actions through the selected WSL project runtime; pins explicit origin preference on runtime open-by-number work item lookups; forwards check-details cancellation without local Git overrides. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. The historical receipt/currentCommand selecting src/main/ssh/ssh-multiplexer-transport-writer.test.ts remains historical and is not a passing result for the reduced current command. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete the complete tail-wait memo suite: its sole current authored callback checks supplied fromTail/blocked field values. Retire all three broader memoization/equivalence/eviction assertion aliases; they are not backed by retained source callbacks. No scan-work budget, sustained output, physical terminal or resource measurement is assigned to this removed test. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1674 retires changed-source assertion and execution scope without equivalent replacement coverage. Existing unchanged-selector receipts remain scoped to their original sources. Historical prior-source oracle: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: The current executable slice injects main-process renderer backlog pressure, then asserts ordinary terminal-wait detection without a joined-tail allocation and with full prompt-family continuity, per-PTY and total pending-output caps, a 1,024-ID projection cap with split-remainder compaction, preserved sequenced-tail metadata, active-pending protection ahead of background trimming, exact provider-generation pause ownership, renderer exit/data ordering, and ACK-gated in-flight bounds. The negotiation seam proves every initial connection and reconnect offers V1, the same-build relay advertises V1 without launch flags, old clients remain token-free, and method-not-found peers fall back without installing source ACK publication. The joined source-intake/multiplex oracle blocks renderer or headless serialization, admits one snapshot-covered span and one trailing span, requires replacement reservation only after the authoritative sequence is known, advances upstream ACK eligibility exactly to the covered boundary after SnapshotEnd, and delivers then ACKs the trailing span through the ordinary live path without cancellation. Source-range ledger and multiplex seams additionally require cumulative byte ACK 40 then 100 for one 100-byte frame to release exact byte credit without early source settlement, admit a contiguous higher-generation recovered token while the prior token remains unsettled, and reject stale stream or source generations without releasing in-flight byte credit. Replacement reservation rolls back every earlier span on a later failure, commit requires the exact frozen transfer states, and an authoritatively reclaimed covered span rejects commit without local trim authority while rollback removes the reservation idempotently. The exit-deadline oracle publishes a source-backed desktop projection, advances a fake clock to cancellation, and requires one projection transfer, proof commit, preparation, and final exit with no provider close, retained obligation, duplicate finalization, or process-lifetime cancellation tombstone; a second controlled proof remains pending across generation close and must never publish final exit, while both paths release the renderer preparation lease exactly once. The model-migration oracle accepts span A, blocks span B's raw emulator completion, begins disconnect migration, and proves no checkpoint or reconnect attach can remain at A; releasing B advances the exact checkpoint to B and projects B once. Its fake-clock arm reaches the 10-second deadline, requires checkpoint-unavailable, one exact model reset, zero retained admission charge and timers, and no effect from late raw settlement. Its two-PTY failure arm rejects one running raw callback under an active exact migration and requires one checkpoint-unavailable result, one model reset, zero provider closes, retained sibling checkpoint and transport liveness, zero charges and timers, and an unaffected other generation; a non-migrating sibling oracle still requires generation close. The intake additionally closes 2,048 sequential provider generations into one exact range, rejects stale events across that range, and preserves an out-of-order live gap until its own close. Direct-SSH contracts additionally prove activation-response settlement before the first source frame, claim-gated projection under same-decoder-turn response/data delivery, exact early-ACK reservation through send settlement or same-token retry, recovery completion visible before lease-held frames transfers those frames only into the private quarantine and admits no output until exact checkpoint-to-recoveryEnd coverage, invalid-checkpoint restore retention through response settlement followed by fresh-token retry and one live source frame, isolated saturated-subscriber eviction with healthy owner liveness, preserved legacy-primary backpressure, one-write ownership across write(false)/drain, control selection ahead of queued ordinary PTY input with FIFO lanes and bounded fairness, liveness rebasing, failed-exit retention through late ACK or exact recovery, response-settlement cancellation authority, token-scoped exit cancellation, exact checkpoint-to-recoveryEnd continuity, stale-owner one-shot retry, no physical PTY teardown on recovery failure, exact provider-generation closure and once-only cleanup when recovery cancellation publication or proof rejects, 16 MiB-plus-1 MiB decoder caps, charged per-PTY/session retention budgets, and eight-wide isolated reattach. The live Electron perf oracle adds hidden-output floods, renderer scheduler queue depth, dropped-output-zero normal scenarios, and active key latency budgets before promotion.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts", "pnpm exec vitest run --config config/vitest.config.ts src/main/runtime/construction/orca-runtime.test.ts", @@ -9370,7 +9354,8 @@ "Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical.", "Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed.", "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed.", - "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed." + "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed.", + "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed." ], "demotionRule": "Cannot promote without metric artifacts and bounded-byte assertions." }, @@ -9606,7 +9591,7 @@ "providers": ["ssh", "remote-runtime", "wsl"], "coveredPlatforms": ["linux"], "coveredProviders": ["ssh", "remote-runtime"], - "coverageNotes": "Audit3315 execution amendment: the reduced two-suite reconnect selector passed on Linux (2 cases). Original source-review statements about no fresh execution describe the earlier static review; this new receipt validates only these two local deterministic unit cases, not the complete gate or physical/runtime/rendered/provider/platform parity. Historical seven-case macOS receipt remains archived verbatim. Audit 3312 retires the ordinary six-pane distinct-live-handle reconnect/resubscription alarm from paired-reconnect-multi-pane-materialization.test.ts. The unchanged parked-pane activation/rematerialization case remains narrower and is not an equivalent replacement. The complete prior seven-contract macOS receipt is archived as historical evidence; it does not validate the reduced selector. No fresh socket, PTY, sleep/wake, provider, platform, mobile, mixed-version or rendered execution is claimed. Audit3096: remove the deleted reconnect-banner selector and its two UI aliases. The reduced recovery-state selector retains only the deadline cancelling pending backoff and explicit retry owning a new epoch alias; healthy-replacement stale-schedule and independent caller-cutoff cancellation aliases retire. Other retained state cases cover reentrant retry, owner settlement and same-epoch parked revival, without replacing those removed alarms. Prior admission-selector retirement, archived July 28 receipt, unrelated selectors/claims/receipts and old gaps remain unchanged; no fresh execution is claimed. Audit3086: retire the destroyed fresh-session fallback settlement/non-destructive reattach alias and its direct selector; retain the detach/attach selector and assertion unchanged. The retained admission cases cover adopt-before-data and explicitly buffered-death admission without a new spawn, without replacing fallback retirement. July 28 evidence was already archived by audit3077; no duplicate retirement or new execution is claimed. Audit2709 removes src/renderer/src/runtime/web-session-tabs-sync/inventory/remote-runtime-session-tabs-inflight.test.ts: intentionally retire waiting out an older inventory and sharing a fresh post-operation request among concurrent callers. No equivalent replacement or fresh execution is claimed. Dated receipts remain historical evidence for their unchanged selectors only; they do not validate the reduced commands or full current gate. Historical prior account (preserved verbatim): Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/ssh/connection/ssh-connection-manager.test.ts: tears a failed startup down quietly, so its error is not replaced by a disconnect; marks a dial that never reached the host, so setup reports SSH’s reason; removes the registration when bulk teardown fails; leaves a connection whose own reconnect is asking the user alone; clears the pool entry when the closed connection is still the registered one; still reuses a connection that is already up; dials a user's connection test. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2052 (sources2049/2050/2051): Remove complete low-value restored-pane-title and pane-visibility-identity suites and their current selectors. Their supplied snapshot/title and nullable identity-comparison alarms are intentionally lost without replacement. Unrelated dated evidenceRuns remain unchanged and do not validate the reduced commands. No fresh gate execution, actual PTY/provider, rendered, remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Delete the terminal-create idempotency forwarding suite and its prospective selector. Preserve2 unrelated receipts without claiming the lost forwarding alarm was replaced. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete complete transport-unmount and queued-initial-CWD helper suites: despite lifecycle-shaped fixtures, the unmount owner returns a Boolean PTY-id predicate, while CWD cases check nullish field selection and callback counts. Lose their detach/destroy/teardown and one-shot-CWD supplied examples. Partial queued-pane-startup loses three identical/empty/nonowner identity examples, retaining four exact one-shot/replacement ownership callbacks. No assertion object maps to these three sources, so unrelated alias objects remain byte-exact. Preserve the current81ed prior removal of pane-split-startup; stale bd948 source metadata snapshots are not replayed. No actual process/PTY startup, transport shutdown, live SSH or rendered recovery is proven by this metadata change. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Delete the sole supplied public-SSH-state spread/sanitized-error projection and its provider-authority-pair alias. Also delete the complete tiny one-shot split-startup wrapper suite: callback-visible temporary field/reset, sequential payload and thrown-sentinel reset alarms disappear. No SSH provider epoch handoff, connection, process startup or actual terminal split was performed by either deleted suite. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1903 removes whole mocked detach, supplied geometry and/or close-request suites selected by this gate. Their pending-CWD/PTY-binding argument/order, fabricated hit-test and close bookkeeping alarms are intentionally lost without replacement. Valid remaining selectors/assertions and unchanged-selector execution receipts retain their original scope. Reduced unit selections have no fresh broad gate, real PTY/provider, rendered focus, or measured-budget validation. Historical prior account (preserved verbatim): Audit1645 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1342 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Audit1301 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Deterministic renderer coverage proves retained native and runtime SSH payloads are admitted through production routes only with valid complete authority, stale cleanup cannot unregister a replacement runtime terminal, direct SSH Git and folder panes keep their PTY bindings through disconnect and reconnect, one authority chain stops after two automatic attempts even when each timeout exceeds the rolling window, rejected acknowledgements mutate no store maps, and one shared exact attempt admits every concurrent split-pane spawn and reattach while preserving the first PTY as the tab fallback. A later sibling failure rotates the tab once, stale callbacks from the prior attempt mutate no state, split remount activity suppression is counted per leaf, primary PTY exit promotes a bound survivor or preserves an empty continuation gap for a late sibling, and primary, non-primary, or null-PTY detach preserves exact authority on both resulting tabs. Intentional pane disposal cancels its settlement timer without breaking StrictMode remount timeout ownership. A mounted remote-runtime terminal survives repeated transport partitions without changing PTY identity. A real encrypted-WebSocket oracle proves a successful reachability probe can replace a pre-ready shared-control socket without rejecting or duplicating the waiting RPC. Direct SSH coordinator tests cover immediate terminal finalization, hydration correction, damping, bounded retry, and telemetry non-interference. Client/server heartbeat tests cover timer suspension, socket generations fence stale callbacks, cold restored-terminal attachment retries, cached pixels remain unhealthy until authoritative replay, automatic retries stop after one minute, manual reconnect preserves the PTY, and pane closure releases recovery UI state. Current-transport recovery clears every error that transport surfaced without clearing or displaying a sibling pane's errors, while transport suppression, pane retention, and the rendered surface are independently bounded. Current macOS Electron journeys against an ephemeral Linux Docker SSH target cover exact-authority repo/worktree hydration, live terminal recovery after disconnect/reconnect, and eager six-terminal remount after renderer reload. A Windows remote-runtime smoke covers reachability and PTY round-trip. Multi-target live fanout, paired-close, WSL, and patched live partition journeys remain gaps. The test-value audit narrowed resolved-worktree-execution-host.test.ts to cache reuse and invalidation. Its direct hydrated-worktree, missing-folder, restored-owner, host-stamp and empty-runtime selection alarms are lost. Other routing and recovery suites do not establish equivalent selector coverage; historical run commands and results remain historical. The test-value audit removes standalone unwritten/released renderer-authority and cross-scope request-count examples. The retained synthetic mirror/store/sidebar and pending-inventory checks do not establish real PTY byte ownership, SSH reconnection or equivalent replacement coverage; historical commands and results remain historical. The test-value audit removes direct join/label ledgers, immediate failure and negative-only stale/removal/partial-authority examples, plus the standalone retained/live event projection. Retained synthetic store/hydration compositions do not establish real SSH continuity, remote authority delivery or equivalent replacement alarms. Historical commands and results remain historical. The renderer scheduling fixture retains in-flight coalescing only. Its immediate publication and negative stale-SSH-cleanup call ledgers were removed; the retained SSH recovery evidence does not replace those alarms. The accepted-handle-events observer suite is removed, including its scoped listener and settlement-release assertions. These observer alarms are lost; retained SSH/transport journeys are not equivalent replacement coverage. The test-value audit removed repeated remote-handle rotation inventories and snapshot-first pending-restart migration examples from terminal-pty-identity-replacement.test.ts, plus direct foreground metadata setter, confirmation, evidence-enum, and prefix/worktree cleanup examples. Those rotation, pending-restart, snapshot-first migration, and foreground scalar-update alarms are lost. The retained replacement case observes pane and tab identities in one subscriber publication; it does not establish equivalent migration, process-evidence, or input-routing coverage. Historical commands and results remain historical. The test-value audit narrows terminals-hydration.test.ts to a real reconnect batch with one measured store publication and hydration-seeded history consumed by a subsequent visit/Back check. Direct split-layout, runtime/folder/floating restoration, dead-leaf focus, cold/known-absent SSH deferral, title, default-tab marker, navigation initialization, and hydrationSucceeded scalar examples are removed; those specific alarms are lost. No actual target-removal transition, live SSH connection, session writer, or persisted restart is exercised by the retained cases. The old target-scoped/snapshot assertion references are retired because the current file does not contain those oracles. Other suites and historical runs do not establish equivalent replacement coverage. Audit1299 removes tests/e2e/terminal-cold-activation-deferral.spec.ts. Private-state/work-budget/assembled-fixture alarms are intentionally lost; retained suites are not equivalent replacement coverage. No new runtime/platform evidence is claimed. Audit1301: src/renderer/src/runtime/remote-runtime-session-tabs-inflight.test.ts reduced. Direct dedupe eviction/count and equal receipt-position sample loses authored alarm. No equivalent replacement coverage is claimed. Audit1318: src/shared/remote-runtime/control/remote-runtime-shared-control-connection.test.ts: Retain real local encrypted WebSocket regressions for pre-ready socket replacement preserving a request, close-before-ready logical subscription cleanup, unrelated RPC surviving another deadline, and pending setup abort leaving a sibling pending. Host server is synthetic and uses production crypto; kept scenarios also retain internal diagnostics/memory assertions unchanged. Remove happy routing, authority-envelope, unknown-ID/log, cleanup enum, binary, keepalive/liveness and production-budget/private socket inventories; intentionally lose those alarms, no physical host authorization, independent crypto or remote provider claim. Reduced selectors need fresh validation; historical full-suite receipts do not validate the changed coverage. No equivalent replacement or physical-provider claim. Audit1342: src/renderer/src/hooks/ipc-events/direct-ssh-connect-reply-routing.test.ts removed. Initial and event-triggered partial-authority hydration fixture alarms disappear. src/renderer/src/components/terminal-pane/terminal-error-accumulation.test.ts removed. Multiline dedup/storm bounds, resource classification/collision, scoped clear/reconnect strip and1000-pane map count alarms disappear. No equivalent replacement coverage is claimed. Audit1396: src/renderer/src/components/terminal-pane/pty-connection-pending-spawn.spec.ts reduced. Lose standalone debug-state retention/log silence alarm. React StrictMode is not mounted; bindings manually simulate remount over mocked transport. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1400: src/main/ipc/ssh/ssh.test.ts reduced. Lose trivial missing target and mocked publication inventory. Preexisting unresolved ../ssh/ssh-port-scanner mock is recorded, not fixed. src/renderer/src/runtime/sync-runtime-graph-scheduling.test.ts removed. Removes whole coalescing call-count suite; no newest-state delivery or replacement claim. src/renderer/src/store/slices/terminal-pty-identity-replacement.test.ts reduced. Removes three no-op/work-count registrations; retains externally observed ownership consistency. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1412: src/main/ssh/ssh-provider-authority.test.ts removed. Loses real AbortController revocation/target isolation/reentrancy outcomes, recreated-target generation-floor/session rollover behavior, stale mutation refusal and settled-registration cleanup alarms. These useful alarms are deliberately lost; no equivalent cancellation coverage is claimed. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1743 removes the standalone line-height clamp/default and pane cwd fallback helper tests (5 authored cases). These trivial-helper alarms are intentionally lost, with no equivalent replacement. Only current commands/testFiles lose the deleted selectors; historical receipts and evidenceRuns remain unchanged and do not validate the reduced commands.", + "coverageNotes": "Audit 3318 reduces runtime-environment-ssh-state.test.ts, intentionally losing metadata membership-pruning/new-target-only-read and other-environment/local-bucket isolation alarms. Preserve generation/failed-full-refresh assertion and unknown-target authoritative-list reread; these narrower cases are not equivalent replacements. The unchanged active two-case Linux reconnect materialization/sidebar receipt and covered platform retain their exact scope and do not validate the reduced SSH state selector. Preserve unrelated SSH assertions, selectors and archived history. No fresh SSH discovery, provider/PTY/network, rendered, mobile, mixed-version, cross-platform or full-gate execution is claimed. Audit3315 execution amendment: the reduced two-suite reconnect selector passed on Linux (2 cases). Original source-review statements about no fresh execution describe the earlier static review; this new receipt validates only these two local deterministic unit cases, not the complete gate or physical/runtime/rendered/provider/platform parity. Historical seven-case macOS receipt remains archived verbatim. Audit 3312 retires the ordinary six-pane distinct-live-handle reconnect/resubscription alarm from paired-reconnect-multi-pane-materialization.test.ts. The unchanged parked-pane activation/rematerialization case remains narrower and is not an equivalent replacement. The complete prior seven-contract macOS receipt is archived as historical evidence; it does not validate the reduced selector. No fresh socket, PTY, sleep/wake, provider, platform, mobile, mixed-version or rendered execution is claimed. Audit3096: remove the deleted reconnect-banner selector and its two UI aliases. The reduced recovery-state selector retains only the deadline cancelling pending backoff and explicit retry owning a new epoch alias; healthy-replacement stale-schedule and independent caller-cutoff cancellation aliases retire. Other retained state cases cover reentrant retry, owner settlement and same-epoch parked revival, without replacing those removed alarms. Prior admission-selector retirement, archived July 28 receipt, unrelated selectors/claims/receipts and old gaps remain unchanged; no fresh execution is claimed. Audit3086: retire the destroyed fresh-session fallback settlement/non-destructive reattach alias and its direct selector; retain the detach/attach selector and assertion unchanged. The retained admission cases cover adopt-before-data and explicitly buffered-death admission without a new spawn, without replacing fallback retirement. July 28 evidence was already archived by audit3077; no duplicate retirement or new execution is claimed. Audit2709 removes src/renderer/src/runtime/web-session-tabs-sync/inventory/remote-runtime-session-tabs-inflight.test.ts: intentionally retire waiting out an older inventory and sharing a fresh post-operation request among concurrent callers. No equivalent replacement or fresh execution is claimed. Dated receipts remain historical evidence for their unchanged selectors only; they do not validate the reduced commands or full current gate. Historical prior account (preserved verbatim): Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/ssh/connection/ssh-connection-manager.test.ts: tears a failed startup down quietly, so its error is not replaced by a disconnect; marks a dial that never reached the host, so setup reports SSH’s reason; removes the registration when bulk teardown fails; leaves a connection whose own reconnect is asking the user alone; clears the pool entry when the closed connection is still the registered one; still reuses a connection that is already up; dials a user's connection test. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2052 (sources2049/2050/2051): Remove complete low-value restored-pane-title and pane-visibility-identity suites and their current selectors. Their supplied snapshot/title and nullable identity-comparison alarms are intentionally lost without replacement. Unrelated dated evidenceRuns remain unchanged and do not validate the reduced commands. No fresh gate execution, actual PTY/provider, rendered, remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Delete the terminal-create idempotency forwarding suite and its prospective selector. Preserve2 unrelated receipts without claiming the lost forwarding alarm was replaced. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete complete transport-unmount and queued-initial-CWD helper suites: despite lifecycle-shaped fixtures, the unmount owner returns a Boolean PTY-id predicate, while CWD cases check nullish field selection and callback counts. Lose their detach/destroy/teardown and one-shot-CWD supplied examples. Partial queued-pane-startup loses three identical/empty/nonowner identity examples, retaining four exact one-shot/replacement ownership callbacks. No assertion object maps to these three sources, so unrelated alias objects remain byte-exact. Preserve the current81ed prior removal of pane-split-startup; stale bd948 source metadata snapshots are not replayed. No actual process/PTY startup, transport shutdown, live SSH or rendered recovery is proven by this metadata change. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Delete the sole supplied public-SSH-state spread/sanitized-error projection and its provider-authority-pair alias. Also delete the complete tiny one-shot split-startup wrapper suite: callback-visible temporary field/reset, sequential payload and thrown-sentinel reset alarms disappear. No SSH provider epoch handoff, connection, process startup or actual terminal split was performed by either deleted suite. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1903 removes whole mocked detach, supplied geometry and/or close-request suites selected by this gate. Their pending-CWD/PTY-binding argument/order, fabricated hit-test and close bookkeeping alarms are intentionally lost without replacement. Valid remaining selectors/assertions and unchanged-selector execution receipts retain their original scope. Reduced unit selections have no fresh broad gate, real PTY/provider, rendered focus, or measured-budget validation. Historical prior account (preserved verbatim): Audit1645 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1342 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Audit1301 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Deterministic renderer coverage proves retained native and runtime SSH payloads are admitted through production routes only with valid complete authority, stale cleanup cannot unregister a replacement runtime terminal, direct SSH Git and folder panes keep their PTY bindings through disconnect and reconnect, one authority chain stops after two automatic attempts even when each timeout exceeds the rolling window, rejected acknowledgements mutate no store maps, and one shared exact attempt admits every concurrent split-pane spawn and reattach while preserving the first PTY as the tab fallback. A later sibling failure rotates the tab once, stale callbacks from the prior attempt mutate no state, split remount activity suppression is counted per leaf, primary PTY exit promotes a bound survivor or preserves an empty continuation gap for a late sibling, and primary, non-primary, or null-PTY detach preserves exact authority on both resulting tabs. Intentional pane disposal cancels its settlement timer without breaking StrictMode remount timeout ownership. A mounted remote-runtime terminal survives repeated transport partitions without changing PTY identity. A real encrypted-WebSocket oracle proves a successful reachability probe can replace a pre-ready shared-control socket without rejecting or duplicating the waiting RPC. Direct SSH coordinator tests cover immediate terminal finalization, hydration correction, damping, bounded retry, and telemetry non-interference. Client/server heartbeat tests cover timer suspension, socket generations fence stale callbacks, cold restored-terminal attachment retries, cached pixels remain unhealthy until authoritative replay, automatic retries stop after one minute, manual reconnect preserves the PTY, and pane closure releases recovery UI state. Current-transport recovery clears every error that transport surfaced without clearing or displaying a sibling pane's errors, while transport suppression, pane retention, and the rendered surface are independently bounded. Current macOS Electron journeys against an ephemeral Linux Docker SSH target cover exact-authority repo/worktree hydration, live terminal recovery after disconnect/reconnect, and eager six-terminal remount after renderer reload. A Windows remote-runtime smoke covers reachability and PTY round-trip. Multi-target live fanout, paired-close, WSL, and patched live partition journeys remain gaps. The test-value audit narrowed resolved-worktree-execution-host.test.ts to cache reuse and invalidation. Its direct hydrated-worktree, missing-folder, restored-owner, host-stamp and empty-runtime selection alarms are lost. Other routing and recovery suites do not establish equivalent selector coverage; historical run commands and results remain historical. The test-value audit removes standalone unwritten/released renderer-authority and cross-scope request-count examples. The retained synthetic mirror/store/sidebar and pending-inventory checks do not establish real PTY byte ownership, SSH reconnection or equivalent replacement coverage; historical commands and results remain historical. The test-value audit removes direct join/label ledgers, immediate failure and negative-only stale/removal/partial-authority examples, plus the standalone retained/live event projection. Retained synthetic store/hydration compositions do not establish real SSH continuity, remote authority delivery or equivalent replacement alarms. Historical commands and results remain historical. The renderer scheduling fixture retains in-flight coalescing only. Its immediate publication and negative stale-SSH-cleanup call ledgers were removed; the retained SSH recovery evidence does not replace those alarms. The accepted-handle-events observer suite is removed, including its scoped listener and settlement-release assertions. These observer alarms are lost; retained SSH/transport journeys are not equivalent replacement coverage. The test-value audit removed repeated remote-handle rotation inventories and snapshot-first pending-restart migration examples from terminal-pty-identity-replacement.test.ts, plus direct foreground metadata setter, confirmation, evidence-enum, and prefix/worktree cleanup examples. Those rotation, pending-restart, snapshot-first migration, and foreground scalar-update alarms are lost. The retained replacement case observes pane and tab identities in one subscriber publication; it does not establish equivalent migration, process-evidence, or input-routing coverage. Historical commands and results remain historical. The test-value audit narrows terminals-hydration.test.ts to a real reconnect batch with one measured store publication and hydration-seeded history consumed by a subsequent visit/Back check. Direct split-layout, runtime/folder/floating restoration, dead-leaf focus, cold/known-absent SSH deferral, title, default-tab marker, navigation initialization, and hydrationSucceeded scalar examples are removed; those specific alarms are lost. No actual target-removal transition, live SSH connection, session writer, or persisted restart is exercised by the retained cases. The old target-scoped/snapshot assertion references are retired because the current file does not contain those oracles. Other suites and historical runs do not establish equivalent replacement coverage. Audit1299 removes tests/e2e/terminal-cold-activation-deferral.spec.ts. Private-state/work-budget/assembled-fixture alarms are intentionally lost; retained suites are not equivalent replacement coverage. No new runtime/platform evidence is claimed. Audit1301: src/renderer/src/runtime/remote-runtime-session-tabs-inflight.test.ts reduced. Direct dedupe eviction/count and equal receipt-position sample loses authored alarm. No equivalent replacement coverage is claimed. Audit1318: src/shared/remote-runtime/control/remote-runtime-shared-control-connection.test.ts: Retain real local encrypted WebSocket regressions for pre-ready socket replacement preserving a request, close-before-ready logical subscription cleanup, unrelated RPC surviving another deadline, and pending setup abort leaving a sibling pending. Host server is synthetic and uses production crypto; kept scenarios also retain internal diagnostics/memory assertions unchanged. Remove happy routing, authority-envelope, unknown-ID/log, cleanup enum, binary, keepalive/liveness and production-budget/private socket inventories; intentionally lose those alarms, no physical host authorization, independent crypto or remote provider claim. Reduced selectors need fresh validation; historical full-suite receipts do not validate the changed coverage. No equivalent replacement or physical-provider claim. Audit1342: src/renderer/src/hooks/ipc-events/direct-ssh-connect-reply-routing.test.ts removed. Initial and event-triggered partial-authority hydration fixture alarms disappear. src/renderer/src/components/terminal-pane/terminal-error-accumulation.test.ts removed. Multiline dedup/storm bounds, resource classification/collision, scoped clear/reconnect strip and1000-pane map count alarms disappear. No equivalent replacement coverage is claimed. Audit1396: src/renderer/src/components/terminal-pane/pty-connection-pending-spawn.spec.ts reduced. Lose standalone debug-state retention/log silence alarm. React StrictMode is not mounted; bindings manually simulate remount over mocked transport. No equivalent replacement coverage is claimed. The existing primary-selector receipt and its claims remain scoped to their unchanged sources; the broad optional command has no new execution result. Audit1400: src/main/ipc/ssh/ssh.test.ts reduced. Lose trivial missing target and mocked publication inventory. Preexisting unresolved ../ssh/ssh-port-scanner mock is recorded, not fixed. src/renderer/src/runtime/sync-runtime-graph-scheduling.test.ts removed. Removes whole coalescing call-count suite; no newest-state delivery or replacement claim. src/renderer/src/store/slices/terminal-pty-identity-replacement.test.ts reduced. Removes three no-op/work-count registrations; retains externally observed ownership consistency. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1412: src/main/ssh/ssh-provider-authority.test.ts removed. Loses real AbortController revocation/target isolation/reentrancy outcomes, recreated-target generation-floor/session rollover behavior, stale mutation refusal and settled-registration cleanup alarms. These useful alarms are deliberately lost; no equivalent cancellation coverage is claimed. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1743 removes the standalone line-height clamp/default and pane cwd fallback helper tests (5 authored cases). These trivial-helper alarms are intentionally lost, with no equivalent replacement. Only current commands/testFiles lose the deleted selectors; historical receipts and evidenceRuns remain unchanged and do not validate the reduced commands.", "motivatingLinks": [ "https://linear.app/stably/issue/STA-3107", "https://github.com/stablyai/orca/pull/12664", @@ -9621,7 +9606,7 @@ "https://github.com/stablyai/orca/issues/12902" ], "invariant": "SSH, WSL, and remote-runtime restore paths must treat provider listing failures and unknown liveness as unknown, not dead, while still avoiding duplicate spawn and clearing expired relay leases exactly once. Direct SSH disconnect, reconnect, and wake must keep PTY bindings and relay identity, and target snapshot hydration/reconnect cannot reset sibling SSH, local, WSL, or runtime-owned state. Every restored remote terminal must preserve its provider PTY identity, including the authoritative incarnation returned by a successful session-ID reattach. After a recoverable partition the same authenticated runtime must reattach the same PTY, reject detached input, apply the latest viewport, and report healthy only after authoritative replay. A successful one-shot reachability probe may replace a pre-ready shared-control socket, but waiting RPCs must continue onto the replacement under their original deadline without duplicate host delivery or retained request bytes. Automatic PTY recovery stops after one bounded minute without a fatal terminal error; a manual reconnect starts a newly fenced epoch against the same PTY, and closed panes retain no recovery UI state. Successful current-transport recovery must clear every stale error surfaced for that pane without clearing or displaying a sibling pane's error; a later genuine failure must remain visible. Each transport retains at most eight suppressed messages, each pane retains at most eight distinct messages, and the rendered surface retains at most 24 lines and 4,000 characters. One capability-gated terminal-create mutation must produce at most one host PTY across an unknown response outcome, remain manually retryable after cutoff, and never let a stale completion replace a newer pane lifecycle. Reconnect must alternate exact activation with authoritative inventory so neither a stale activation response nor an activation failure can strand or retire a pane, and activating a parked surface whose persisted binding was already retired must respawn it rather than report a changed owner after signalling its exit.", - "oracle": "Audit 3312 retires the ordinary six-pane distinct-live-handle reconnect/resubscription alarm from paired-reconnect-multi-pane-materialization.test.ts. The unchanged parked-pane activation/rematerialization case remains narrower and is not an equivalent replacement. The complete prior seven-contract macOS receipt is archived as historical evidence; it does not validate the reduced selector. No fresh socket, PTY, sleep/wake, provider, platform, mobile, mixed-version or rendered execution is claimed. Audit3096: remove the deleted reconnect-banner selector and its two UI aliases. The reduced recovery-state selector retains only the deadline cancelling pending backoff and explicit retry owning a new epoch alias; healthy-replacement stale-schedule and independent caller-cutoff cancellation aliases retire. Other retained state cases cover reentrant retry, owner settlement and same-epoch parked revival, without replacing those removed alarms. Prior admission-selector retirement, archived July 28 receipt, unrelated selectors/claims/receipts and old gaps remain unchanged; no fresh execution is claimed. Audit3086: retire the destroyed fresh-session fallback settlement/non-destructive reattach alias and its direct selector; retain the detach/attach selector and assertion unchanged. The retained admission cases cover adopt-before-data and explicitly buffered-death admission without a new spawn, without replacing fallback retirement. July 28 evidence was already archived by audit3077; no duplicate retirement or new execution is claimed. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/ssh/connection/ssh-connection-manager.test.ts: tears a failed startup down quietly, so its error is not replaced by a disconnect; marks a dial that never reached the host, so setup reports SSH’s reason; removes the registration when bulk teardown fails; leaves a connection whose own reconnect is asking the user alone; clears the pool entry when the closed connection is still the registered one; still reuses a connection that is already up; dials a user's connection test. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2052 (sources2049/2050/2051): Remove complete low-value restored-pane-title and pane-visibility-identity suites and their current selectors. Their supplied snapshot/title and nullable identity-comparison alarms are intentionally lost without replacement. Unrelated dated evidenceRuns remain unchanged and do not validate the reduced commands. No fresh gate execution, actual PTY/provider, rendered, remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Delete the terminal-create idempotency forwarding suite and its prospective selector. Preserve2 unrelated receipts without claiming the lost forwarding alarm was replaced. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete complete transport-unmount and queued-initial-CWD helper suites: despite lifecycle-shaped fixtures, the unmount owner returns a Boolean PTY-id predicate, while CWD cases check nullish field selection and callback counts. Lose their detach/destroy/teardown and one-shot-CWD supplied examples. Partial queued-pane-startup loses three identical/empty/nonowner identity examples, retaining four exact one-shot/replacement ownership callbacks. No assertion object maps to these three sources, so unrelated alias objects remain byte-exact. Preserve the current81ed prior removal of pane-split-startup; stale bd948 source metadata snapshots are not replayed. No actual process/PTY startup, transport shutdown, live SSH or rendered recovery is proven by this metadata change. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Delete the sole supplied public-SSH-state spread/sanitized-error projection and its provider-authority-pair alias. Also delete the complete tiny one-shot split-startup wrapper suite: callback-visible temporary field/reset, sequential payload and thrown-sentinel reset alarms disappear. No SSH provider epoch handoff, connection, process startup or actual terminal split was performed by either deleted suite. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1903 removes whole mocked detach, supplied geometry and/or close-request suites selected by this gate. Their pending-CWD/PTY-binding argument/order, fabricated hit-test and close bookkeeping alarms are intentionally lost without replacement. Valid remaining selectors/assertions and unchanged-selector execution receipts retain their original scope. Reduced unit selections have no fresh broad gate, real PTY/provider, rendered focus, or measured-budget validation. Historical prior account (preserved verbatim): Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): Deterministic tests cover bounded stale-handle replacement, suspended heartbeat clocks, cold and established subscription failure, ten partition/recovery cycles, automatic-recovery cutoff, manual reconnect, and direct SSH binding retention. They assert SSH state changes leave PTY bindings in place, relay identity survives, and another target/local/WSL/runtime panes remain byte-identical through target snapshot hydration and reconnect. Both remount callbacks consume split-count activity suppression, intentional dispose emits no failure/timeout, and a same-attempt StrictMode remount still owns one timeout. Hydration clears an untrusted PTY hint without clearing its current pending owner, healthy current-authority bindings suppress correction, hydration finalizes once, and reconnect emits no paired close lifecycle. A provider-level session-ID reattach returns an incarnation, then a legacy exit without an incarnation must resolve to that returned identity rather than minting a fallback identity. The shared-control oracle withholds the first encrypted ready frame, starts one RPC, triggers the successful-probe refresh, then requires exactly two client connections, one host request, a successful response, zero pending calls, and zero retained request bytes. Tests also assert one unsubscribe per remote-runtime epoch, observable recovery phases, stable PTY identity, resumed snapshot/output/input, no healthy state before replay, no retry or input after cutoff, a new manual epoch against the same PTY, quiet recovery UI with an explicit Reconnect action, pane-close state cleanup, one stable create mutation id, old-runtime no-retry behavior, cross-process PTY adoption, and bounded in-flight coordination. A focused recovery oracle surfaces two distinct failures through one transport, completes authoritative replay, and requires both matching clear callbacks; pane-state tests require only the active pane's error to render, matching-pane recovery to preserve sibling and unrelated errors, dismissal to re-admit later failures, and distinct storms to retain only the newest eight messages within 24 lines and 4,000 characters. The retained store replacement assertion observes one atomic pane/tab subscriber publication; repeated rotations and snapshot-first pending-restart migration no longer have their direct store examples.", + "oracle": "Audit 3318 reduces runtime-environment-ssh-state.test.ts, intentionally losing metadata membership-pruning/new-target-only-read and other-environment/local-bucket isolation alarms. Preserve generation/failed-full-refresh assertion and unknown-target authoritative-list reread; these narrower cases are not equivalent replacements. The unchanged active two-case Linux reconnect materialization/sidebar receipt and covered platform retain their exact scope and do not validate the reduced SSH state selector. Preserve unrelated SSH assertions, selectors and archived history. No fresh SSH discovery, provider/PTY/network, rendered, mobile, mixed-version, cross-platform or full-gate execution is claimed. Audit 3312 retires the ordinary six-pane distinct-live-handle reconnect/resubscription alarm from paired-reconnect-multi-pane-materialization.test.ts. The unchanged parked-pane activation/rematerialization case remains narrower and is not an equivalent replacement. The complete prior seven-contract macOS receipt is archived as historical evidence; it does not validate the reduced selector. No fresh socket, PTY, sleep/wake, provider, platform, mobile, mixed-version or rendered execution is claimed. Audit3096: remove the deleted reconnect-banner selector and its two UI aliases. The reduced recovery-state selector retains only the deadline cancelling pending backoff and explicit retry owning a new epoch alias; healthy-replacement stale-schedule and independent caller-cutoff cancellation aliases retire. Other retained state cases cover reentrant retry, owner settlement and same-epoch parked revival, without replacing those removed alarms. Prior admission-selector retirement, archived July 28 receipt, unrelated selectors/claims/receipts and old gaps remain unchanged; no fresh execution is claimed. Audit3086: retire the destroyed fresh-session fallback settlement/non-destructive reattach alias and its direct selector; retain the detach/attach selector and assertion unchanged. The retained admission cases cover adopt-before-data and explicitly buffered-death admission without a new spawn, without replacing fallback retirement. July 28 evidence was already archived by audit3077; no duplicate retirement or new execution is claimed. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/ssh/connection/ssh-connection-manager.test.ts: tears a failed startup down quietly, so its error is not replaced by a disconnect; marks a dial that never reached the host, so setup reports SSH’s reason; removes the registration when bulk teardown fails; leaves a connection whose own reconnect is asking the user alone; clears the pool entry when the closed connection is still the registered one; still reuses a connection that is already up; dials a user's connection test. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2052 (sources2049/2050/2051): Remove complete low-value restored-pane-title and pane-visibility-identity suites and their current selectors. Their supplied snapshot/title and nullable identity-comparison alarms are intentionally lost without replacement. Unrelated dated evidenceRuns remain unchanged and do not validate the reduced commands. No fresh gate execution, actual PTY/provider, rendered, remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1983: Delete the terminal-create idempotency forwarding suite and its prospective selector. Preserve2 unrelated receipts without claiming the lost forwarding alarm was replaced. Changed-source prior results do not validate reduced-source bytes. No fresh execution, equivalent replacement, red/green, measured budget, soak, provider/PTY, rendered, live remote/mobile or cross-platform result is claimed. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Delete complete transport-unmount and queued-initial-CWD helper suites: despite lifecycle-shaped fixtures, the unmount owner returns a Boolean PTY-id predicate, while CWD cases check nullish field selection and callback counts. Lose their detach/destroy/teardown and one-shot-CWD supplied examples. Partial queued-pane-startup loses three identical/empty/nonowner identity examples, retaining four exact one-shot/replacement ownership callbacks. No assertion object maps to these three sources, so unrelated alias objects remain byte-exact. Preserve the current81ed prior removal of pane-split-startup; stale bd948 source metadata snapshots are not replayed. No actual process/PTY startup, transport shutdown, live SSH or rendered recovery is proven by this metadata change. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Delete the sole supplied public-SSH-state spread/sanitized-error projection and its provider-authority-pair alias. Also delete the complete tiny one-shot split-startup wrapper suite: callback-visible temporary field/reset, sequential payload and thrown-sentinel reset alarms disappear. No SSH provider epoch handoff, connection, process startup or actual terminal split was performed by either deleted suite. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1903 removes whole mocked detach, supplied geometry and/or close-request suites selected by this gate. Their pending-CWD/PTY-binding argument/order, fabricated hit-test and close bookkeeping alarms are intentionally lost without replacement. Valid remaining selectors/assertions and unchanged-selector execution receipts retain their original scope. Reduced unit selections have no fresh broad gate, real PTY/provider, rendered focus, or measured-budget validation. Historical prior account (preserved verbatim): Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): Deterministic tests cover bounded stale-handle replacement, suspended heartbeat clocks, cold and established subscription failure, ten partition/recovery cycles, automatic-recovery cutoff, manual reconnect, and direct SSH binding retention. They assert SSH state changes leave PTY bindings in place, relay identity survives, and another target/local/WSL/runtime panes remain byte-identical through target snapshot hydration and reconnect. Both remount callbacks consume split-count activity suppression, intentional dispose emits no failure/timeout, and a same-attempt StrictMode remount still owns one timeout. Hydration clears an untrusted PTY hint without clearing its current pending owner, healthy current-authority bindings suppress correction, hydration finalizes once, and reconnect emits no paired close lifecycle. A provider-level session-ID reattach returns an incarnation, then a legacy exit without an incarnation must resolve to that returned identity rather than minting a fallback identity. The shared-control oracle withholds the first encrypted ready frame, starts one RPC, triggers the successful-probe refresh, then requires exactly two client connections, one host request, a successful response, zero pending calls, and zero retained request bytes. Tests also assert one unsubscribe per remote-runtime epoch, observable recovery phases, stable PTY identity, resumed snapshot/output/input, no healthy state before replay, no retry or input after cutoff, a new manual epoch against the same PTY, quiet recovery UI with an explicit Reconnect action, pane-close state cleanup, one stable create mutation id, old-runtime no-retry behavior, cross-process PTY adoption, and bounded in-flight coordination. A focused recovery oracle surfaces two distinct failures through one transport, completes authoritative replay, and requires both matching clear callbacks; pane-state tests require only the active pane's error to render, matching-pane recovery to preserve sibling and unrelated errors, dismissal to re-admit later failures, and distinct storms to retain only the newest eight messages within 24 lines and 4,000 characters. The retained store replacement assertion observes one atomic pane/tab subscriber publication; repeated rotations and snapshot-first pending-restart migration no longer have their direct store examples.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts src/renderer/src/runtime/web-session-tabs-sync/agent-status/paired-reconnect-sidebar-agent-count.test.ts --reporter=dot", "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal/background-terminal-worktree-mount.test.ts src/renderer/src/components/terminal-pane/mounting/queued-pane-startup.test.ts src/renderer/src/components/terminal-pane/parking/one-shot-park-replay.test.ts src/renderer/src/components/terminal-pane/pty-connection/tests/pty-connection-spawn-and-startup.test.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-host-surface-replacement.test.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-stream-reconnect.test.ts src/renderer/src/store/terminals/terminal-pty-identity-replacement.test.ts", @@ -9767,9 +9752,8 @@ { "file": "src/renderer/src/runtime/hosts/ssh/runtime-environment-ssh-state.test.ts", "assertions": [ - "membership refresh prunes removed targets and reads only newly joined targets", "held replies from an old generation cannot replace a newer generation; failed full refresh preserves live state behind queued metadata", - "positive environment and local state remain isolated; an unknown target triggers an authoritative list reread" + "an unknown target triggers an authoritative list reread" ] }, { @@ -9882,7 +9866,8 @@ "Module-layout wave2 intentionally retires complete prohibited owning registrations in src/renderer/src/lib/resolved-worktree-execution-host.test.ts. Lost alarms: Both explicit and legacy workspace owners resolve to env-a initially. Replacing the repo catalog refreshes legacy ownership to env-b. A conflicting local repo row makes ownership ambiguous and returns null. Repeated unrelated updates perform zero id reads, an internal work-budget alarm removed. No equivalent replacement coverage or reduced-selector pass is claimed.", "Audit 3312 retires the ordinary six-pane distinct-live-handle reconnect/resubscription alarm from paired-reconnect-multi-pane-materialization.test.ts. The unchanged parked-pane activation/rematerialization case remains narrower and is not an equivalent replacement. The complete prior seven-contract macOS receipt is archived as historical evidence; it does not validate the reduced selector. No fresh socket, PTY, sleep/wake, provider, platform, mobile, mixed-version or rendered execution is claimed.", "Audit 3312 retires this complete dated seven-contract receipt because paired-reconnect-multi-pane-materialization.test.ts was reduced. Original command, currentCommand, result, counts, date and duration remain historical and do not validate the reduced selector. Original receipt verbatim:\n{\n \"date\": \"2026-08-23\",\n \"runner\": \"local\",\n \"platform\": \"macos\",\n \"command\": \"pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/paired-reconnect-multi-pane-materialization.test.ts src/renderer/src/runtime/paired-reconnect-sidebar-agent-count.test.ts --reporter=dot\",\n \"result\": \"passed\",\n \"durationSeconds\": 2.54,\n \"summary\": \"Seven deterministic STA-3107 contracts passed across six-pane multiplex rematerialization and client-owned sidebar-row retention after sleep/wake.\",\n \"currentCommand\": \"pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/paired-reconnect-multi-pane-materialization.test.ts src/renderer/src/runtime/web-session-tabs-sync/agent-status/paired-reconnect-sidebar-agent-count.test.ts --reporter=dot\"\n }", - "Audit3315 execution amendment: the reduced two-suite reconnect selector passed on Linux (2 cases). Original source-review statements about no fresh execution describe the earlier static review; this new receipt validates only these two local deterministic unit cases, not the complete gate or physical/runtime/rendered/provider/platform parity. Historical seven-case macOS receipt remains archived verbatim." + "Audit3315 execution amendment: the reduced two-suite reconnect selector passed on Linux (2 cases). Original source-review statements about no fresh execution describe the earlier static review; this new receipt validates only these two local deterministic unit cases, not the complete gate or physical/runtime/rendered/provider/platform parity. Historical seven-case macOS receipt remains archived verbatim.", + "Audit 3318 reduces runtime-environment-ssh-state.test.ts, intentionally losing metadata membership-pruning/new-target-only-read and other-environment/local-bucket isolation alarms. Preserve generation/failed-full-refresh assertion and unknown-target authoritative-list reread; these narrower cases are not equivalent replacements. The unchanged active two-case Linux reconnect materialization/sidebar receipt and covered platform retain their exact scope and do not validate the reduced SSH state selector. Preserve unrelated SSH assertions, selectors and archived history. No fresh SSH discovery, provider/PTY/network, rendered, mobile, mixed-version, cross-platform or full-gate execution is claimed." ], "demotionRule": "Cannot promote if provider failure can close panes or if the oracle is screenshot-only." }, @@ -12742,14 +12727,14 @@ "providers": ["local", "remote-runtime", "ssh"], "coveredPlatforms": ["macos", "linux"], "coveredProviders": ["local", "remote-runtime", "ssh"], - "coverageNotes": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2056 (sources2054/2055): Intentionally retire these low-value alarms without replacement: src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts: delivers remote file changes to listeners for the same worktree. src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/mobile-session-tabs-part-14.spec'). src/main/runtime/orca-runtime-tests/terminal-side-effect-facts-part-03.spec.ts: arms the stale-title timer for a seeded working title; arms the Command Code scrape from the noted spawn command; prefers the tracked title over a stale renderer lastTitle in the hydration seed; writes recent output into the current runtime output map; uses a later title handler replacement with the runtime receiver; calls the current injected listener with the runtime receiver; keeps optional reconciliation lazy after its callback is removed. src/main/runtime/orca-runtime-tests/mobile-session-tabs-part-14.spec.ts: lists them, so a paired client can mirror a migrated editor; closes one by retiring it from the host session; keeps an unsaved draft unless the close is forced; shares aggregate work but leaves a replacement pending promise alone; lets a targeted refresh finish while an aggregate refresh waits; refuses a close when the store lacks %s before the guard; keeps a later missing %s on the original caught failure path. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Exact retired assertionRefs, preserved as historical objects only: {\n \"file\": \"src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts\",\n \"assertions\": [\n \"terminal watcher error and setup failure evict the shared entry before retrying callbacks run\",\n \"an old stream end cannot evict the replacement stream, which still delivers to its listener\",\n \"two same-worktree listeners share a synthetic subscription until the final stop unsubscribes and sends unwatch\"\n ]\n } Historical prior account (preserved verbatim): Audit1973: Delete the standalone mocked watcher cleanup suite and its entire sole-file prospective command. Retire the removed file-command transport-abort alias; preserve the independent cancellation suite and final-overflow/error/end cleanup callback, and13 unrelated receipt objects. Whole original changed dated receipts are retired, never narrowed or transferred. This metadata review claims no fresh execution, equivalent replacement, provider/PTY, live remote/mobile, rendered, cross-platform, soak or measured-budget result. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Partial runtime files-watch suite retains controlled Windows close-event waiting, shared-client survival, late physical-exit ownership/deadline cleanup and shutdown joining. Its combined Windows close/deadline/late-exit alias remains backed by unchanged callbacks. Retire native-error-close, pre-publication setup ownership and failed-delete rearm aliases; associated delegation, failure forwarding, removal routing and recovery examples are lost. These tests inject watch/subscription seams and do not observe an actual native watcher or physical process. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Partial runtime watcher-host suite loses mocked option/event/overflow/resubscription forwarding, supplied teardown-error propagation, aborted-signal/no-call and private ten-thousand-waiter-count inventories. Retire the 10,000-caller bound and failed-recovery-subscriber-ending assertion aliases. Keep same-root owner cancellation, pre-ready replacement, held root-close join, stale-generation recovery delivery and initial failed-root ownership mappings backed by retained real-host callbacks with injected subscription/error/physical-exit promises. No actual native watcher, process termination or resource bound was observed. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1683 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1645 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1627 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1623 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1516 removes package-electron-runtime-contract.test.mjs, a copied release-workflow platform and marker matrix. Its workflow wiring alarm is intentionally retired; no actual packaged watcher delivery was exercised and no equivalent replacement is claimed. Unaffected dated execution receipts retain their original scope; changed optional selectors have no fresh full-gate execution result. This audit removes the paired-web stopped-during-outage and late-ready direct-branch examples. The retained pre-ready cancellation, failed-unwatch retry and replay-continuity cases are not equivalent replacements for those ordering alarms. Current web-runtime-client.test.ts assertions below name only its retained pre-ready callback case; replay cases remain in their separate suite. Historical commands and results remain historical. Deterministic tests cover one shared healthy child per runtime process plus at most four bounded fault-quarantine children, native desktop, paired-runtime, and WSL snapshot processes sharing the same eight-physical-child reservation with typed serialized event-driven capacity recovery, including recursive re-wait when crash recovery reclaims an announced slot, one quarantine attempt per watch lifetime, generation-scoped cancellation, bounded termination with removable deadline waiters, replacement-crawl errors that cannot be blessed by a late readiness ack, pre-ready paired-web cancellation registered before an unbounded capacity wait, already-resolved and late physical-exit retry of a rejected paired-web teardown, physical desktop-install and local/SSH PTY teardown, stale-generation rejection, per-child stat bounds, final-overflow RPC delivery, and renderer eviction before terminal callbacks can retry. Destructive local and SSH removal fences both ID-derived and resolved-cwd terminal roots, including sibling-root/cwd combinations with reverse admission rollback, closes descendant relay watches before parent deletion, enumerates authoritative provider/cwd ownership, falls back to daemon spawn cwd before OSC 7, and keeps daemon/relay immediate shutdown pending until native exit or a bounded fail-closed error; a dead relay PID discovered during attach also settles concurrent shutdown before its stale entry is reaped, while Windows runtime watcher deadlines retain one removable physical-close waiter, treat Node's error-close path as positive physical-exit proof, and clear root ownership on late close. The relay policy test maps a standard repository's base, Git common directory, and worktree roots to one healthy supervisor, then proves a shared-child failure recovers those roots in separate quarantine supervisors with overflow and resumed events. Abortable capacity, quarantine, runtime-root, relay, SSH, and relay pre-install setup waits attach one reaction to each shared promise and explicitly remove 10,000 cancelled caller closures while one anchor remains. Built-entry macOS harnesses kill the desktop/runtime and SSH relay watcher children, then require automatic resubscription and later events while the host process and relay PTY survive. SSH bundles the child boundary beside relay.js, requires both artifacts for install completeness, and preserves registration-owned same-root cancellation. The test-value audit corrects the active renderer watch command/file reference to the existing runtime-file-client-watch.test.ts, whose three cases are retained unchanged. They observe synthetic subscription retry/fanout and unsubscribe/unwatch calls, not actual preload/main bookkeeping or physical watcher-process cleanup. The separate read/preview client suite loses forwarding, supplied-error and negative-only examples, retaining binary detection and Unicode SVG decoding. Historical evidence commands/results are unchanged. Test-value audit1293 removes src/main/providers/local-pty/local-pty-provider-session-inventory.test.ts (2 authored cases). Their isolated assertions no longer run; surviving tests are not claimed as equivalent coverage. Fresh execution against the candidate SHA is required. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1412: src/main/workspaces/removal/watcher-removal-gate.test.ts reduced. Loses exact removal-error compatibility with renderer predicate under shared message constants. Kept public refusal outcomes do not establish that renderer classification. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result.", + "coverageNotes": "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2056 (sources2054/2055): Intentionally retire these low-value alarms without replacement: src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts: delivers remote file changes to listeners for the same worktree. src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/mobile-session-tabs-part-14.spec'). src/main/runtime/orca-runtime-tests/terminal-side-effect-facts-part-03.spec.ts: arms the stale-title timer for a seeded working title; arms the Command Code scrape from the noted spawn command; prefers the tracked title over a stale renderer lastTitle in the hydration seed; writes recent output into the current runtime output map; uses a later title handler replacement with the runtime receiver; calls the current injected listener with the runtime receiver; keeps optional reconciliation lazy after its callback is removed. src/main/runtime/orca-runtime-tests/mobile-session-tabs-part-14.spec.ts: lists them, so a paired client can mirror a migrated editor; closes one by retiring it from the host session; keeps an unsaved draft unless the close is forced; shares aggregate work but leaves a replacement pending promise alone; lets a targeted refresh finish while an aggregate refresh waits; refuses a close when the store lacks %s before the guard; keeps a later missing %s on the original caught failure path. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Exact retired assertionRefs, preserved as historical objects only: {\n \"file\": \"src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts\",\n \"assertions\": [\n \"terminal watcher error and setup failure evict the shared entry before retrying callbacks run\",\n \"an old stream end cannot evict the replacement stream, which still delivers to its listener\",\n \"two same-worktree listeners share a synthetic subscription until the final stop unsubscribes and sends unwatch\"\n ]\n } Historical prior account (preserved verbatim): Audit1973: Delete the standalone mocked watcher cleanup suite and its entire sole-file prospective command. Retire the removed file-command transport-abort alias; preserve the independent cancellation suite and final-overflow/error/end cleanup callback, and13 unrelated receipt objects. Whole original changed dated receipts are retired, never narrowed or transferred. This metadata review claims no fresh execution, equivalent replacement, provider/PTY, live remote/mobile, rendered, cross-platform, soak or measured-budget result. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Partial runtime files-watch suite retains controlled Windows close-event waiting, shared-client survival, late physical-exit ownership/deadline cleanup and shutdown joining. Its combined Windows close/deadline/late-exit alias remains backed by unchanged callbacks. Retire native-error-close, pre-publication setup ownership and failed-delete rearm aliases; associated delegation, failure forwarding, removal routing and recovery examples are lost. These tests inject watch/subscription seams and do not observe an actual native watcher or physical process. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Partial runtime watcher-host suite loses mocked option/event/overflow/resubscription forwarding, supplied teardown-error propagation, aborted-signal/no-call and private ten-thousand-waiter-count inventories. Retire the 10,000-caller bound and failed-recovery-subscriber-ending assertion aliases. Keep same-root owner cancellation, pre-ready replacement, held root-close join, stale-generation recovery delivery and initial failed-root ownership mappings backed by retained real-host callbacks with injected subscription/error/physical-exit promises. No actual native watcher, process termination or resource bound was observed. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1683 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1645 removes low-value test projections, mock wiring, private-state assertions, inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained public behavior and refusal boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1640 removes low-value IPC, SSH, workspace and transcript-watcher tests without equivalent replacement coverage. Per-source reviews record retained public refusal, cancellation, execution-host and real-filesystem outcomes with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1634 removes low-value IPC, filesystem and runtime projections, argument inventories and duplicate examples without equivalent replacement coverage. Per-source reviews record retained cancellation, ownership, authorization, filesystem and SQLite boundaries with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1627 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1623 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1516 removes package-electron-runtime-contract.test.mjs, a copied release-workflow platform and marker matrix. Its workflow wiring alarm is intentionally retired; no actual packaged watcher delivery was exercised and no equivalent replacement is claimed. Unaffected dated execution receipts retain their original scope; changed optional selectors have no fresh full-gate execution result. This audit removes the paired-web stopped-during-outage and late-ready direct-branch examples. The retained pre-ready cancellation, failed-unwatch retry and replay-continuity cases are not equivalent replacements for those ordering alarms. Current web-runtime-client.test.ts assertions below name only its retained pre-ready callback case; replay cases remain in their separate suite. Historical commands and results remain historical. Deterministic tests cover one shared healthy child per runtime process plus at most four bounded fault-quarantine children, native desktop, paired-runtime, and WSL snapshot processes sharing the same eight-physical-child reservation with typed serialized event-driven capacity recovery, including recursive re-wait when crash recovery reclaims an announced slot, one quarantine attempt per watch lifetime, generation-scoped cancellation, bounded termination with removable deadline waiters, replacement-crawl errors that cannot be blessed by a late readiness ack, pre-ready paired-web cancellation registered before an unbounded capacity wait, already-resolved and late physical-exit retry of a rejected paired-web teardown, physical desktop-install and local/SSH PTY teardown, stale-generation rejection, per-child stat bounds, final-overflow RPC delivery, and renderer eviction before terminal callbacks can retry. Destructive local and SSH removal fences both ID-derived and resolved-cwd terminal roots, including sibling-root/cwd combinations with reverse admission rollback, closes descendant relay watches before parent deletion, enumerates authoritative provider/cwd ownership, falls back to daemon spawn cwd before OSC 7, and keeps daemon/relay immediate shutdown pending until native exit or a bounded fail-closed error; a dead relay PID discovered during attach also settles concurrent shutdown before its stale entry is reaped, while Windows runtime watcher deadlines retain one removable physical-close waiter, treat Node's error-close path as positive physical-exit proof, and clear root ownership on late close. The relay policy test maps a standard repository's base, Git common directory, and worktree roots to one healthy supervisor, then proves a shared-child failure recovers those roots in separate quarantine supervisors with overflow and resumed events. Abortable capacity, quarantine, runtime-root, relay, SSH, and relay pre-install setup waits attach one reaction to each shared promise and explicitly remove 10,000 cancelled caller closures while one anchor remains. Built-entry macOS harnesses kill the desktop/runtime and SSH relay watcher children, then require automatic resubscription and later events while the host process and relay PTY survive. SSH bundles the child boundary beside relay.js, requires both artifacts for install completeness, and preserves registration-owned same-root cancellation. The test-value audit corrects the active renderer watch command/file reference to the existing runtime-file-client-watch.test.ts, whose three cases are retained unchanged. They observe synthetic subscription retry/fanout and unsubscribe/unwatch calls, not actual preload/main bookkeeping or physical watcher-process cleanup. The separate read/preview client suite loses forwarding, supplied-error and negative-only examples, retaining binary detection and Unicode SVG decoding. Historical evidence commands/results are unchanged. Test-value audit1293 removes src/main/providers/local-pty/local-pty-provider-session-inventory.test.ts (2 authored cases). Their isolated assertions no longer run; surviving tests are not claimed as equivalent coverage. Fresh execution against the candidate SHA is required. Audit1312 removes low-value imported specs from aggregate test drivers; old whole-driver execution receipts are archived. Retained registrations are unchanged. No equivalent replacement coverage is claimed. Audit1383: src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts reduced. Lose command-list/receipt/default/background activation matrices whose provider spawn returns canned IDs. Retained test proves actual tab allocation separation, not terminal startup. No equivalent replacement coverage is claimed. Existing exact unaffected receipts and their platform/provider/budget claims remain scoped to unchanged selectors. Changed optional selectors have no new execution result. Audit1401: src/main/runtime/orca-runtime-tests/mobile-session-tabs-duplicate-pty-repair.spec.ts removed. Loses listing-on-already-repaired-data alarm; it never established runtime repair invocation. src/main/runtime/orca-runtime-tests/mobile-session-tabs-split-agent-tab.spec.ts removed. Loses actual mobile split sibling shell versus recorded-agent resume command alarms; no equivalent coverage asserted. src/main/runtime/orca-runtime-tests/worktree-native-chat-setup.spec.ts removed. Loses separate setup/startup tab result in that manually assembled path. src/main/runtime/orca-runtime.test.ts reduced. Driver registration/import of a whole-deleted selected spec is removed; directly authored cases remain unchanged. No extra authored removal credit. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result. Audit1412: src/main/workspaces/removal/watcher-removal-gate.test.ts reduced. Loses exact removal-error compatibility with renderer predicate under shared message constants. Kept public refusal outcomes do not establish that renderer classification. No equivalent replacement coverage is claimed. Existing exact unaffected receipts remain unchanged and apply only to their unchanged selectors. Changed optional selectors have no new execution result.", "motivatingLinks": [ "https://github.com/stablyai/orca/issues/5308", "https://github.com/stablyai/orca/issues/7547", "https://github.com/stablyai/orca/issues/8212" ], "invariant": "Local runtime and SSH relay roots share one watcher child per host process while healthy and use at most four quarantine children after shard failure; native desktop, paired-runtime, and direct WSL snapshot children consume the same process-wide eight-child reservation and retry transient exhaustion without permanently disabling a root. Cancellation remains owned by its child generation and settles only after acknowledgement or bounded physical exit; paired-web retains cleanup ownership and can cancel before readiness even while capacity is exhausted, retries a failed unwatch after already-resolved or late physical exit without duplicating stream termination, replacements never overlap a still-retiring generation, and aborted or timed-out callers detach immediately even while queued behind same-root relay teardown. Destructive local and SSH cleanup holds one containment-aware watcher/terminal removal fence through descendant watcher and PTY physical exit plus parent Git/filesystem deletion; daemon and relay kill requests retain ownership until the actual subprocess exit callback or an equivalent positive dead-PID proof and propagate rejected signals or a bounded failure otherwise. A crash-fused shared shard moves affected roots into the bounded quarantine pool, a second fused quarantine ends those watch lifetimes, interrupted setup emits overflow after recovery, metadata work stays bounded, and the SSH relay, PTYs, Git, and other filesystem operations never load the watcher native addon or share its fault boundary.", - "oracle": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2056 (sources2054/2055): Intentionally retire these low-value alarms without replacement: src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts: delivers remote file changes to listeners for the same worktree. src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/mobile-session-tabs-part-14.spec'). src/main/runtime/orca-runtime-tests/terminal-side-effect-facts-part-03.spec.ts: arms the stale-title timer for a seeded working title; arms the Command Code scrape from the noted spawn command; prefers the tracked title over a stale renderer lastTitle in the hydration seed; writes recent output into the current runtime output map; uses a later title handler replacement with the runtime receiver; calls the current injected listener with the runtime receiver; keeps optional reconciliation lazy after its callback is removed. src/main/runtime/orca-runtime-tests/mobile-session-tabs-part-14.spec.ts: lists them, so a paired client can mirror a migrated editor; closes one by retiring it from the host session; keeps an unsaved draft unless the close is forced; shares aggregate work but leaves a replacement pending promise alone; lets a targeted refresh finish while an aggregate refresh waits; refuses a close when the store lacks %s before the guard; keeps a later missing %s on the original caught failure path. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Exact retired assertionRefs, preserved as historical objects only: {\n \"file\": \"src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts\",\n \"assertions\": [\n \"terminal watcher error and setup failure evict the shared entry before retrying callbacks run\",\n \"an old stream end cannot evict the replacement stream, which still delivers to its listener\",\n \"two same-worktree listeners share a synthetic subscription until the final stop unsubscribes and sends unwatch\"\n ]\n } Historical prior account (preserved verbatim): Audit1973: Delete the standalone mocked watcher cleanup suite and its entire sole-file prospective command. Retire the removed file-command transport-abort alias; preserve the independent cancellation suite and final-overflow/error/end cleanup callback, and13 unrelated receipt objects. Whole original changed dated receipts are retired, never narrowed or transferred. This metadata review claims no fresh execution, equivalent replacement, provider/PTY, live remote/mobile, rendered, cross-platform, soak or measured-budget result. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Partial runtime files-watch suite retains controlled Windows close-event waiting, shared-client survival, late physical-exit ownership/deadline cleanup and shutdown joining. Its combined Windows close/deadline/late-exit alias remains backed by unchanged callbacks. Retire native-error-close, pre-publication setup ownership and failed-delete rearm aliases; associated delegation, failure forwarding, removal routing and recovery examples are lost. These tests inject watch/subscription seams and do not observe an actual native watcher or physical process. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Partial runtime watcher-host suite loses mocked option/event/overflow/resubscription forwarding, supplied teardown-error propagation, aborted-signal/no-call and private ten-thousand-waiter-count inventories. Retire the 10,000-caller bound and failed-recovery-subscriber-ending assertion aliases. Keep same-root owner cancellation, pre-ready replacement, held root-close join, stale-generation recovery delivery and initial failed-root ownership mappings backed by retained real-host callbacks with injected subscription/error/physical-exit promises. No actual native watcher, process termination or resource bound was observed. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1683 retires changed-source assertion and execution scope without equivalent replacement coverage. Existing unchanged-selector receipts remain scoped to their original sources. Historical prior-source oracle: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Fill the global watcher-child reservation, require native desktop, paired-runtime, and WSL setup to retain typed transient capacity and retry after one serialized release, then cancel pending desktop and paired-client watches and require no retry. Hold one live anchor while aborting 10,000 callers at the capacity, quarantine-assignment, runtime-root, relay-setup, and SSH-registration layers; require exactly one explicit waiter and one base-promise reaction to remain, then zero after settlement. Hold an old relay unsubscribe, queue one same-root setup leader and one anchor, then abort 10,000 joiners; require every cancellation to settle before teardown releases, exactly one anchor waiter, and only one later physical install. Advance the former paired-web callback deadline while capacity remains full, require cancellation ownership to stay live, publish the server token before ready, and require files.unwatch to abort setup before a later slot release can install a native root. Reject the first ready-watch unwatch with a physical-exit promise, resolve that exit later, and require a fresh physical unwatch attempt, registry release, and exactly one stream end. Fill one healthy shared child with multiple roots, map a standard SSH repository's base, Git common directory, and worktree roots to one relay supervisor, then inject IPC disconnect-before-exit, queued and active cancellation, operations during termination, kill-without-exit, repeated unkillable generations, initial and recovery timeouts, and shard failure. During local and SSH deletion, attempt a late terminal spawn whose worktree identity and cwd name different roots, enumerate a non-prefix remote PTY by authoritative worktree ownership, withhold daemon/relay exit callbacks, make attach positively prove a concurrent-shutdown PID dead, and make the provider unavailable; require every acquired admission to roll back on rejection, dead-PID proof to settle physical exit before reaping, timeout to retain ownership, and Git/filesystem deletion never to start. Require bounded watcher process count, physical-exit settlement, healthy-root recovery, stale-generation rejection, overflow plus later events, and fail-closed destructive cleanup. Verify relay build hashes and remote install probes require both relay.js and relay-watcher.js. Launch the built relay with a real PTY and Parcel watch, kill only the watcher child, and require the same relay PID and PTY to answer before a replacement child delivers overflow and a post-crash event.", + "oracle": "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2090 (sources2088/2089): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2086 (sources2082/2083): Current reduced suites intentionally retire the alarms listed in knownGaps; prior dated execution and replacement claims do not validate changed source bytes or selectors. Historical prior account (preserved verbatim): Audit2076 (sources2073/2074): Intentionally retire these low-value alarms without replacement: src/main/runtime/orca-runtime-tests/terminal-handles-part-02.spec.ts: recovers a moved SSH pane through the tab it now sits in, not the one its lease froze; reports a plain expired orphan; does not report a pane whose only recent lease was superseded; does not report a pane whose only recent lease had its relay id recycled; picks the eligible orphan over a superseded predecessor that matches first; collects the same tabs the per-tab reader reports, in one sweep of the leases; reports no candidate when no lease names any of the worktree tabs. src/main/runtime/orca-runtime-tests/runtime-availability.spec.ts: starts unavailable with no authoritative window; reports the configured Windows terminal shell on status; reports floating workspace availability from settings on status; polls floating tabs with targeted PTY liveness and no repo/provider inventory; hydrates persisted tabs when the store cannot report repos; advertises browser screencast only when a renderer window is available; routes mobile Codex reset consumption through the account mutation coordinator; maps a definite pre-provider rejection into an authoritative current snapshot. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Frozen current literal-import graph reaches these changed test bytes; unresolved local and computed edges remain debt, and current reachability is not historical execution or equivalent replacement evidence. Historical prior account (preserved verbatim): Audit2056 (sources2054/2055): Intentionally retire these low-value alarms without replacement: src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts: delivers remote file changes to listeners for the same worktree. src/main/runtime/orca-runtime.test.ts: ordinary registration cleanup removes only await import('./orca-runtime-tests/mobile-session-tabs-part-14.spec'). src/main/runtime/orca-runtime-tests/terminal-side-effect-facts-part-03.spec.ts: arms the stale-title timer for a seeded working title; arms the Command Code scrape from the noted spawn command; prefers the tracked title over a stale renderer lastTitle in the hydration seed; writes recent output into the current runtime output map; uses a later title handler replacement with the runtime receiver; calls the current injected listener with the runtime receiver; keeps optional reconciliation lazy after its callback is removed. src/main/runtime/orca-runtime-tests/mobile-session-tabs-part-14.spec.ts: lists them, so a paired client can mirror a migrated editor; closes one by retiring it from the host session; keeps an unsaved draft unless the close is forced; shares aggregate work but leaves a replacement pending promise alone; lets a targeted refresh finish while an aggregate refresh waits; refuses a close when the store lacks %s before the guard; keeps a later missing %s on the original caught failure path. Existing unrelated evidenceRuns retain their original scope; unchanged dated receipt objects do not validate changed source bytes or reduced selectors. No fresh gate execution or historical import-closure equivalence is claimed. Exact retired assertionRefs, preserved as historical objects only: {\n \"file\": \"src/renderer/src/runtime/files/watching/runtime-file-client-watch.test.ts\",\n \"assertions\": [\n \"terminal watcher error and setup failure evict the shared entry before retrying callbacks run\",\n \"an old stream end cannot evict the replacement stream, which still delivers to its listener\",\n \"two same-worktree listeners share a synthetic subscription until the final stop unsubscribes and sends unwatch\"\n ]\n } Historical prior account (preserved verbatim): Audit1973: Delete the standalone mocked watcher cleanup suite and its entire sole-file prospective command. Retire the removed file-command transport-abort alias; preserve the independent cancellation suite and final-overflow/error/end cleanup callback, and13 unrelated receipt objects. Whole original changed dated receipts are retired, never narrowed or transferred. This metadata review claims no fresh execution, equivalent replacement, provider/PTY, live remote/mobile, rendered, cross-platform, soak or measured-budget result. Historical prior account (preserved verbatim): Audit1931 (sources1928/1929): Partial runtime files-watch suite retains controlled Windows close-event waiting, shared-client survival, late physical-exit ownership/deadline cleanup and shutdown joining. Its combined Windows close/deadline/late-exit alias remains backed by unchanged callbacks. Retire native-error-close, pre-publication setup ownership and failed-delete rearm aliases; associated delegation, failure forwarding, removal routing and recovery examples are lost. These tests inject watch/subscription seams and do not observe an actual native watcher or physical process. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1925 (sources1922/1923): Partial runtime watcher-host suite loses mocked option/event/overflow/resubscription forwarding, supplied teardown-error propagation, aborted-signal/no-call and private ten-thousand-waiter-count inventories. Retire the 10,000-caller bound and failed-recovery-subscriber-ending assertion aliases. Keep same-root owner cancellation, pre-ready replacement, held root-close join, stale-generation recovery delivery and initial failed-root ownership mappings backed by retained real-host callbacks with injected subscription/error/physical-exit promises. No actual native watcher, process termination or resource bound was observed. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1683 retires changed-source assertion and execution scope without equivalent replacement coverage. Existing unchanged-selector receipts remain scoped to their original sources. Historical prior-source oracle: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Existing exact unchanged-selector receipts remain the only current execution evidence. Changed source selectors have no fresh execution result. Historical prior-source oracle account: Fill the global watcher-child reservation, require native desktop, paired-runtime, and WSL setup to retain typed transient capacity and retry after one serialized release, then cancel pending desktop and paired-client watches and require no retry. Hold one live anchor while aborting 10,000 callers at the capacity, quarantine-assignment, runtime-root, relay-setup, and SSH-registration layers; require exactly one explicit waiter and one base-promise reaction to remain, then zero after settlement. Hold an old relay unsubscribe, queue one same-root setup leader and one anchor, then abort 10,000 joiners; require every cancellation to settle before teardown releases, exactly one anchor waiter, and only one later physical install. Advance the former paired-web callback deadline while capacity remains full, require cancellation ownership to stay live, publish the server token before ready, and require files.unwatch to abort setup before a later slot release can install a native root. Reject the first ready-watch unwatch with a physical-exit promise, resolve that exit later, and require a fresh physical unwatch attempt, registry release, and exactly one stream end. Fill one healthy shared child with multiple roots, map a standard SSH repository's base, Git common directory, and worktree roots to one relay supervisor, then inject IPC disconnect-before-exit, queued and active cancellation, operations during termination, kill-without-exit, repeated unkillable generations, initial and recovery timeouts, and shard failure. During local and SSH deletion, attempt a late terminal spawn whose worktree identity and cwd name different roots, enumerate a non-prefix remote PTY by authoritative worktree ownership, withhold daemon/relay exit callbacks, make attach positively prove a concurrent-shutdown PID dead, and make the provider unavailable; require every acquired admission to roll back on rejection, dead-PID proof to settle physical exit before reaping, timeout to retain ownership, and Git/filesystem deletion never to start. Require bounded watcher process count, physical-exit settlement, healthy-root recovery, stale-generation rejection, overflow plus later events, and fail-closed destructive cleanup. Verify relay build hashes and remote install probes require both relay.js and relay-watcher.js. Launch the built relay with a real PTY and Parcel watch, kill only the watcher child, and require the same relay PID and PTY to answer before a replacement child delivers overflow and a post-crash event.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-unsubscribe-timeout.test.ts --reporter=dot", "pnpm exec vitest run --config config/vitest.config.ts src/main/ipc/filesystem/watchers/filesystem-watcher.test.ts src/main/ipc/filesystem/watchers/filesystem-watcher-native-capacity.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-child-registry.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-child-launch.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-process.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-supervisor-capacity.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-disconnect-termination.test.ts src/main/ipc/filesystem/watcher-process/runtime-watcher-pending-assignment.test.ts src/main/ipc/filesystem/watcher-process/parcel-watcher-event-delivery.test.ts src/main/ipc/filesystem/watchers/filesystem-watcher-local-unsubscribe.test.ts src/main/ipc/filesystem/watchers/filesystem-watcher-wsl.test.ts src/main/ipc/worktrees-delete-pty-teardown.test.ts src/main/ipc/worktrees-orphan-directory-cleanup.test.ts src/main/ipc/worktrees-remove-preflight.test.ts src/main/workspaces/removal/watcher-removal-gate.test.ts src/main/runtime/files/watch/orca-runtime-files-watch.test.ts src/main/runtime/workspaces/removal/worktree-teardown.test.ts src/main/runtime/construction/orca-runtime.test.ts src/main/runtime/rpc/methods/files/files.test.ts src/main/runtime/rpc/methods/files/files-watch-cancellation.test.ts src/renderer/src/web/web-runtime-client.test.ts src/shared/cross-platform-path.test.ts src/shared/physical-exit-tracker.test.ts src/shared/promise-settlement-waiters.test.ts src/main/ipc/pty/pty-spawn-cwd-fallback.test.ts src/main/providers/local-pty/local-pty-provider-shutdown.test.ts src/main/daemon/providers/adapter/testing/daemon-pty-adapter.test.ts src/main/daemon/pty-subprocess/testing/pty-subprocess.test.ts src/main/daemon/sessions/testing/session.test.ts src/main/daemon/terminal-host/testing/terminal-host.test.ts src/main/daemon/terminal-host/testing/terminal-host-session-reaping-leak.test.ts --reporter=dot", @@ -13189,7 +13174,8 @@ "Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical.", "Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed.", "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed.", - "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed." + "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed.", + "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed." ], "demotionRule": "Demote or quarantine if the fault harness flakes without a product or harness bug, if healthy operation exceeds one runtime watcher child, if total physical operation exceeds eight children including retiring generations, if quarantine children outlive their roots or repeat after fusing, if event delivery becomes unbounded, or if metadata/stat work returns to the serve process." }, @@ -13965,7 +13951,7 @@ "providers": ["local-daemon", "ssh-daemon", "paired-runtime"], "coveredPlatforms": ["macos"], "coveredProviders": ["local-daemon", "paired-runtime"], - "coverageNotes": "Audit1939 (sources1934/1935): Pruned cold-parking hook and parked-watcher suites retain local timer/selection/split and exit/cancel/disposal progression. Lose supplied capability/eligibility/topology/admission and duplicate state examples recorded below. No assertion objects map to these sources here; preserve independent actual E2E receipt scopes without transferring their results to reduced units. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1623 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1342 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Deterministic policy and multiplex tests separate renderer parking from authoritative stream liveness, while runtime-graph tests cover exact parked leaf, pane runtime ID, title, multi-pane active-leaf, disposal behavior, and queued split routing to an exact cold-parked tab. A real Electron journey runs the shipped dev CLI against the app's isolated profile, proves the exact parked target alone remounts under a bounded lease, completes before the historical 10-second timeout without stealing active tab or focus, re-parks, then reveals two stable-identity panes with independent keyboard/output round trips. Focused mirror-recovery tests preserve a verified binding across pending-handle snapshots, quarantine positive PTY-identity mismatches, accept authoritative ready replacement or removal, and fence recovery to the pairing revision. Daemon attach-only tests prove a replacement runtime re-registers one active session and one history writer. Headed cold-park and headless runtime-restart paired journeys prove rendered output, input, resize convergence, process identity, authoritative close, checkpoint continuity, and post-restart history append against real daemon PTYs. The parked-leaf graph fixture now retains only PTY exit during the park handoff gap. Its standalone exact leaf/title, multi-pane active-leaf and disposal assertions were removed. web-runtime-session.test.ts no longer supplies the eager worktree-switch/post-create recovery assertion. Those alarms are lost; the remaining cold-park and recovery journeys do not establish equivalent replacement coverage. Recovery references retain old-host-to-new-tab mirroring, adoption/list recovery, blocked local ownership changes, coalescing/overtaking/global concurrency, adoption-conflict serialization and pane-resolution retries. Stable unsupported/malformed adoption tables, supplied inventory/topology/identity and verified-binding shape examples were removed. Their separate alarms are lost; remaining compositions are not equivalent coverage. Historical executed commands and results remain historical. The separate Unsubscribe-to-raw-subscriber-veto-clear sample was removed. Its exact raw-presence transition alarm is lost; remaining fake-provider streaming/attachment checks and real paired journeys do not establish equivalent replacement coverage. The deterministic second subscription is a second live connection, not a disconnect/reconnect proof. Audit1292: The entire orphan-inventory retry suite is removed. Unchanged-snapshot retry for temporarily absent, pending or attached PTYs, live-binding preservation, and consecutive-absence confirmation reset after a ready frame are lost. Mock adoption/list behavior and assembled snapshot handling did not establish actual orphan adoption or equivalent replacement coverage. Audit1342: src/renderer/src/runtime/web-session-terminal-orphan-recovery-regressions.test.ts reduced. Exact global RPC concurrency4/counts for fabricated worktrees. No equivalent replacement coverage is claimed.", + "coverageNotes": "Audit 3318 reduces web-session-terminal-pending-handle-recovery.test.ts, intentionally retiring disconnected retry, two-authoritative-absence retirement, explicit close-record short circuit, older-host disconnected absence, positive published-presence protection and unpublished-frame no-count alarms. Remove only the disconnected-retry assertion alias; preserve transient-retry assertion and both unchanged active Electron receipts. The remaining absence-reset interleaving is narrower, not an equivalent replacement for these six removed cases. Historical receipts retain their original scope. No fresh runtime, host/PTY/network, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit1939 (sources1934/1935): Pruned cold-parking hook and parked-watcher suites retain local timer/selection/split and exit/cancel/disposal progression. Lose supplied capability/eligibility/topology/admission and duplicate state examples recorded below. No assertion objects map to these sources here; preserve independent actual E2E receipt scopes without transferring their results to reduced units. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): Audit1623 removes low-value mocked daemon projections, scalar tables and duplicate terminal examples without equivalent replacement coverage. Per-source reviews record retained terminal content preservation, process ownership refusal and recovery boundaries together with injected/native/remote limits. Affected historical receipts are retired; gates without active current evidence are downgraded. Historical scope before this audit: Audit1342 changes referenced sources. Retained unchanged-source receipts apply only to their unchanged commands; the reduced selectors have no new execution result. Historical coverage account (not a current claim for changed sources): Deterministic policy and multiplex tests separate renderer parking from authoritative stream liveness, while runtime-graph tests cover exact parked leaf, pane runtime ID, title, multi-pane active-leaf, disposal behavior, and queued split routing to an exact cold-parked tab. A real Electron journey runs the shipped dev CLI against the app's isolated profile, proves the exact parked target alone remounts under a bounded lease, completes before the historical 10-second timeout without stealing active tab or focus, re-parks, then reveals two stable-identity panes with independent keyboard/output round trips. Focused mirror-recovery tests preserve a verified binding across pending-handle snapshots, quarantine positive PTY-identity mismatches, accept authoritative ready replacement or removal, and fence recovery to the pairing revision. Daemon attach-only tests prove a replacement runtime re-registers one active session and one history writer. Headed cold-park and headless runtime-restart paired journeys prove rendered output, input, resize convergence, process identity, authoritative close, checkpoint continuity, and post-restart history append against real daemon PTYs. The parked-leaf graph fixture now retains only PTY exit during the park handoff gap. Its standalone exact leaf/title, multi-pane active-leaf and disposal assertions were removed. web-runtime-session.test.ts no longer supplies the eager worktree-switch/post-create recovery assertion. Those alarms are lost; the remaining cold-park and recovery journeys do not establish equivalent replacement coverage. Recovery references retain old-host-to-new-tab mirroring, adoption/list recovery, blocked local ownership changes, coalescing/overtaking/global concurrency, adoption-conflict serialization and pane-resolution retries. Stable unsupported/malformed adoption tables, supplied inventory/topology/identity and verified-binding shape examples were removed. Their separate alarms are lost; remaining compositions are not equivalent coverage. Historical executed commands and results remain historical. The separate Unsubscribe-to-raw-subscriber-veto-clear sample was removed. Its exact raw-presence transition alarm is lost; remaining fake-provider streaming/attachment checks and real paired journeys do not establish equivalent replacement coverage. The deterministic second subscription is a second live connection, not a disconnect/reconnect proof. Audit1292: The entire orphan-inventory retry suite is removed. Unchanged-snapshot retry for temporarily absent, pending or attached PTYs, live-binding preservation, and consecutive-absence confirmation reset after a ready frame are lost. Mock adoption/list behavior and assembled snapshot handling did not establish actual orphan adoption or equivalent replacement coverage. Audit1342: src/renderer/src/runtime/web-session-terminal-orphan-recovery-regressions.test.ts reduced. Exact global RPC concurrency4/counts for fabricated worktrees. No equivalent replacement coverage is claimed.", "motivatingLinks": [ "https://linear.app/stably/issue/STA-2854", "https://github.com/stablyai/orca/pull/15514", @@ -13973,7 +13959,7 @@ "https://github.com/stablyai/orca/issues/17297" ], "invariant": "Cold parking or restarting a host renderer/runtime never retires a live PTY's runtime-graph leaf, erases a paired viewer's last verified leaf binding merely because the host temporarily publishes pending-handle, or drops the surviving PTY's history writer. A split request for a cold-parked tab replays against its stable source leaf after exact-tab remount without switching workspaces or focusing the tab. The paired stream, input, resize, pane identity, multi-pane routing, checkpoint, and output log converge to the authoritative ready handle; the binding clears only after positive replacement, mismatch, or two consecutive authoritative absence observations without an intervening ready surface.", - "oracle": "Audit1939 (sources1934/1935): Pruned cold-parking hook and parked-watcher suites retain local timer/selection/split and exit/cancel/disposal progression. Lose supplied capability/eligibility/topology/admission and duplicate state examples recorded below. No assertion objects map to these sources here; preserve independent actual E2E receipt scopes without transferring their results to reduced units. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): Cold-park a host-owned pane while a paired client actively views it, require the host manager to unmount, then type through the client and require the same PTY to receive and echo the token without any disconnected sample. Restart a real pinned-port headless runtime while its daemon PTY survives, observe a pending-handle snapshot, require the viewer binding never to become empty, then require rendered output, focused keyboard input, a post-recovery window resize, the pre-restart checkpoint, and post-restart output-log append to reach the same PTY and process before authoritative close clears the binding. Invoke the real `orca-dev terminal split` against a cold-parked renderer-owned tab, require only that tab to acquire and release the existing background-mount lease, complete before the historical timeout, preserve the active worktree/tab/focus, then reveal two rendered panes and type through each while retaining the stable source leaf and PTY. Deterministically require synchronous stable-leaf replay after lifecycle registration and fail closed for a missing or stale leaf. Separately publish multi-pane parked leaves with exact pane runtime IDs and require per-PTY disposal to remove only the retired leaf.", + "oracle": "Audit 3318 reduces web-session-terminal-pending-handle-recovery.test.ts, intentionally retiring disconnected retry, two-authoritative-absence retirement, explicit close-record short circuit, older-host disconnected absence, positive published-presence protection and unpublished-frame no-count alarms. Remove only the disconnected-retry assertion alias; preserve transient-retry assertion and both unchanged active Electron receipts. The remaining absence-reset interleaving is narrower, not an equivalent replacement for these six removed cases. Historical receipts retain their original scope. No fresh runtime, host/PTY/network, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit1939 (sources1934/1935): Pruned cold-parking hook and parked-watcher suites retain local timer/selection/split and exit/cancel/disposal progression. Lose supplied capability/eligibility/topology/admission and duplicate state examples recorded below. No assertion objects map to these sources here; preserve independent actual E2E receipt scopes without transferring their results to reduced units. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Historical prior account (preserved verbatim): The retained unchanged-source receipts are the only current execution evidence; reduced-source assertions await validation. Historical oracle account (not a current claim for changed sources): Cold-park a host-owned pane while a paired client actively views it, require the host manager to unmount, then type through the client and require the same PTY to receive and echo the token without any disconnected sample. Restart a real pinned-port headless runtime while its daemon PTY survives, observe a pending-handle snapshot, require the viewer binding never to become empty, then require rendered output, focused keyboard input, a post-recovery window resize, the pre-restart checkpoint, and post-restart output-log append to reach the same PTY and process before authoritative close clears the binding. Invoke the real `orca-dev terminal split` against a cold-parked renderer-owned tab, require only that tab to acquire and release the existing background-mount lease, complete before the historical timeout, preserve the active worktree/tab/focus, then reveal two rendered panes and type through each while retaining the stable source leaf and PTY. Deterministically require synchronous stable-leaf replay after lifecycle registration and fail closed for a missing or stale leaf. Separately publish multi-pane parked leaves with exact pane runtime IDs and require per-PTY disposal to remove only the retired leaf.", "commands": [ "pnpm exec vitest run --config config/vitest.config.ts src/renderer/src/components/terminal-pane/parking/output-watchers/terminal-parked-tab-watchers.test.ts src/renderer/src/components/terminal-pane/layout/split/terminal-pane-split-request-routing.test.ts src/renderer/src/components/terminal-pane/parking/lifecycle/use-terminal-tab-cold-parking.test.ts src/renderer/src/hooks/ipc-events/terminal-ui-routing-ipc-bridge-split.test.ts src/main/runtime/terminal/layout/orca-runtime-terminal-split-authority.test.ts src/renderer/src/runtime/sync-runtime-graph/sync-runtime-graph-parked-leaf.test.ts src/renderer/src/runtime/hosted-terminals/recovery/web-session-terminal-orphan-mixed-version.test.ts src/renderer/src/runtime/hosted-terminals/recovery/web-session-terminal-orphan-recovery.test.ts src/renderer/src/runtime/hosted-terminals/recovery/web-session-terminal-orphan-recovery-regressions.test.ts src/renderer/src/runtime/hosted-terminals/recovery/adoption/web-session-terminal-orphan-recovery-adoption-regressions.test.ts src/renderer/src/runtime/hosted-terminals/recovery/panes/web-session-terminal-pending-handle-recovery.test.ts tests/e2e/host-cold-park-remote-subscriber.unit.test.ts --reporter=dot", "pnpm exec vitest run --config config/vitest.config.ts src/main/daemon/providers/adapter/testing/daemon-pty-adapter-cold-restore-reanchor.test.ts src/main/daemon/providers/adapter/testing/daemon-pty-adapter-session-adoption.test.ts src/main/daemon/providers/adapter/testing/daemon-pty-adapter-protocol-compatibility.test.ts src/main/daemon/providers/adapter/testing/daemon-pty-adapter-history-recovery.test.ts --reporter=dot", @@ -14069,8 +14055,7 @@ { "file": "src/renderer/src/runtime/hosted-terminals/recovery/panes/web-session-terminal-pending-handle-recovery.test.ts", "assertions": [ - "retries a transient pane-resolution failure for an unchanged replayed snapshot", - "retries a disconnected pane resolution for an unchanged replayed snapshot" + "retries a transient pane-resolution failure for an unchanged replayed snapshot" ] } ], @@ -14132,7 +14117,8 @@ "Audit1342 historical platform/provider scope before source changes: {\"coveredPlatforms\":[\"macos\",\"linux\"],\"coveredProviders\":[\"local-daemon\",\"paired-runtime\"]}. Remaining provider scope applies only to unchanged-source receipts.", "Audit1342 historical assertion inventory for deleted/reduced sources; not current validated assertions:\n[\n {\n \"file\": \"src/renderer/src/runtime/web-session-terminal-orphan-recovery-regressions.test.ts\",\n \"assertions\": [\n \"coalesces degraded frames to one active and one latest queued recovery\",\n \"lets a newer fully ready frame overtake an in-flight degraded recovery\",\n \"limits recovery RPC concurrency globally across worktrees\"\n ]\n }\n]", "Audit1939 (sources1934/1935): Pruned cold-parking hook and parked-watcher suites retain local timer/selection/split and exit/cancel/disposal progression. Lose supplied capability/eligibility/topology/admission and duplicate state examples recorded below. No assertion objects map to these sources here; preserve independent actual E2E receipt scopes without transferring their results to reduced units. Preserve meaningful remaining selectors/assertions and unchanged dated receipts within their original scope. Reduced selectors have no fresh gate execution evidence in this metadata packet. No equivalent replacement, physical watcher/PTY/provider, live SSH/remote/mobile, rendered, cross-platform, soak or measured-budget result is claimed. Exact retired authored alarms: 'parks paired-runtime tabs only when their exact host advertises restore'; 'skips parking scans for capability-equivalent status writes and scans membership changes'; 'keeps eviction-exempt tabs mounted when the worktree is force-parked'; 'resolves eviction exemptions once per force-park input change'; 'does not exempt a tab from a truncated malformed pane key'; 'never starts a watcher for a PTY that exited into the park handoff gap'; 'collapses the exited leaf out of the stored layout when deferring'; 'retires launch/title hints when the launch-owning parked leaf exits'; 'keeps launch ownership when only a parked shell sibling exits'; 'rejects ordinary parking for a preserved daemon with a lossy snapshot'; 'rejects a capture containing a PTY without snapshot backing'; 'accepts layout-derived candidates when the capture is stale'; 'rejects layout-derived candidates missing a leaf PTY binding'.", - "Terminal/runtime defining-owner cleanup retires complete parking timer, lifecycle, mocked close/disposal and render/flip/mount work-count cases under repository test policy. The retained watcher source has only two literal-oracle coverage decisions; the retained cold-parking hook source has only the layout-PTY exemption invalidation case. Lost alarms include cooldown/measure/pin wakeups, split exit/close attribution and cancellation, watcher disposal/restart, flood timer rearm, worktree verdict damping, and destroyed-transport cold restore. No equivalent replacement coverage or execution pass for the reduced cases is claimed; prior dated receipts retain their original scope." + "Terminal/runtime defining-owner cleanup retires complete parking timer, lifecycle, mocked close/disposal and render/flip/mount work-count cases under repository test policy. The retained watcher source has only two literal-oracle coverage decisions; the retained cold-parking hook source has only the layout-PTY exemption invalidation case. Lost alarms include cooldown/measure/pin wakeups, split exit/close attribution and cancellation, watcher disposal/restart, flood timer rearm, worktree verdict damping, and destroyed-transport cold restore. No equivalent replacement coverage or execution pass for the reduced cases is claimed; prior dated receipts retain their original scope.", + "Audit 3318 reduces web-session-terminal-pending-handle-recovery.test.ts, intentionally retiring disconnected retry, two-authoritative-absence retirement, explicit close-record short circuit, older-host disconnected absence, positive published-presence protection and unpublished-frame no-count alarms. Remove only the disconnected-retry assertion alias; preserve transient-retry assertion and both unchanged active Electron receipts. The remaining absence-reset interleaving is narrower, not an equivalent replacement for these six removed cases. Historical receipts retain their original scope. No fresh runtime, host/PTY/network, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed." ], "demotionRule": "Demote if a live parked/restarting PTY loses its exact graph leaf or verified viewer binding, a retired PTY remains published, multi-pane identity drifts, or either routed client round trip flakes without a diagnosed cause." }, @@ -14434,10 +14420,10 @@ "providers": ["paired-runtime"], "coveredPlatforms": ["linux"], "coveredProviders": ["paired-runtime"], - "coverageNotes": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit1579 removes workflow/manifest/regex/private binary-header and diagnostic metric inventories, fixture selectors, injected forwarding/platform/signature/process examples, unproven overlap and internal churn counters without equivalent replacement coverage. Retains actual Git trust/diffs; filesystem hash/ELF-region/native discovery; child CLI module repair; compiler/lock/publication/rollback and synthetic Electron install/cache ownership; executable shell quoting/exit; controlled notary/timer cleanup; local patched extractor file/AST freshness; real Chromium/WebKit CSP/sandbox/HTTPS image/referrer/link/preservation controls under parsed Android headers. Five unchanged setup-pnpm/VM request-cleanup definitions reviewed and not rerun. GitHub, installer/archive, OS metric, client, native delegate, compressor, builder and Apple boundaries injected or absent where recorded. No actual native PTY, shipped full/native app, real Actions/cache/npm signature authority, downloaded Electron archive, real objdump/linker/native CLI/AppImage, signing/NSIS/notary/upload/updater or live freeze proof. Historical scope before this audit: Published Linux 1.4.188 desktop against current source in both directions; scheduled weekly and manually runnable. No required PR check.", + "coverageNotes": "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit1579 removes workflow/manifest/regex/private binary-header and diagnostic metric inventories, fixture selectors, injected forwarding/platform/signature/process examples, unproven overlap and internal churn counters without equivalent replacement coverage. Retains actual Git trust/diffs; filesystem hash/ELF-region/native discovery; child CLI module repair; compiler/lock/publication/rollback and synthetic Electron install/cache ownership; executable shell quoting/exit; controlled notary/timer cleanup; local patched extractor file/AST freshness; real Chromium/WebKit CSP/sandbox/HTTPS image/referrer/link/preservation controls under parsed Android headers. Five unchanged setup-pnpm/VM request-cleanup definitions reviewed and not rerun. GitHub, installer/archive, OS metric, client, native delegate, compressor, builder and Apple boundaries injected or absent where recorded. No actual native PTY, shipped full/native app, real Actions/cache/npm signature authority, downloaded Electron archive, real objdump/linker/native CLI/AppImage, signing/NSIS/notary/upload/updater or live freeze proof. Historical scope before this audit: Published Linux 1.4.188 desktop against current source in both directions; scheduled weekly and manually runnable. No required PR check.", "motivatingLinks": ["https://github.com/stablyai/orca/actions/runs/34069063016"], "invariant": "A paired client and host without client-hosted browser capabilities retain server-hosted browser placement across supported version skew.", - "oracle": "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Require both existing named browser placement scenarios to pass three times with one attempt, zero skips, zero failures, and no report errors.", + "oracle": "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed. Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed. Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed. Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed. Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical. Audit 3273 removes 14 authored declarations and 19 expanded scenarios across four whole suites and eight partial suites. Retired alarms cover saved-local cancellation/layout ownership, ambiguous-owner warning reset, folder attachment routing, VM false-capability bypass, abandoned reveal warmup cleanup and viewport claim deduplication/rearming, quiet unrelated remote-removal refusal, startup-command one-shot consumption, repository/floating recovery-ledger rows, fake-React pinned-scroll intent, split-ratio state identity/subscriber suppression and fractional change, cross-host project-group join refusal, three PR request-lifetime permutations, upstream-status preservation on error, and unresolved Git-owner refusal. These regression checks are intentionally lost under the test-value audit; retained scenarios are not equivalent replacements. Historical coverage receipts remain historical, including gate35 subscriber claims. Audit 3256 removes these specific alarms from the broad root unit commands: skill dialog active-share selection, close and pending-share handoff; simple keyed-record deletion; marker live-region insertion, simple landing and nine baseline policy examples; General search/control navigation inventory; orphan-count cache bootstrap; seven Codex code classifications and HTTP 429 warning mapping; two skill unread-root verdict examples; peer-mail payload and conversation-ID rejection; and the GitHub organization TABLE text-field optimistic rollback/failed-envelope example. These include real behaviors; retained tests do not provide equivalent coverage for every removed scenario. Mobile losses are recorded separately and are not coverage of these root commands. Audit3246 retires NewRowsMarker reveal/no-repeat and pane-scoped beforeinput checks; skill-owner option availability/count/dedup/order; six styled Input/Textarea IME forwarding rows; the static Codex/OpenCode smart-key inventory; tab-map AST ratchet diagnostic fixtures; child-roster work-budget and retained-aggregate alarms; malformed persisted floating-cwd fallback/no-crash/empty exact-root checks; the unanswered-host empty-list discriminator; the direct invalidated-cache commit-false assertion; one redundant digit-keyup IME row; cached non-supported Claude title-generator rows; and the duplicate late naming-result singleton. These root broad commands lose those specific alarms without equivalent replacement. Distinct retained owner/state checks have their own bounded claims; the real scheduler still rejects stale response data, while the direct false-return alarm is retired. Audit3231 retires supported-host restore-pane key mapping, closed/open creation-sidebar toggle visibility, unrelated scope-error/null classification and quick-submit reveal/no-create checks. These broad commands lose those real narrow alarms; no replacement coverage is claimed. Audit3219 removes the synthetic Cmd+J extra-source suite; this broad command no longer checks its match/tie placement and selection, announced count, empty-query cap, or workspace hiding with retained recent tabs. No replacement coverage is claimed. Audit3209 composes approved sources3202/3206: deliberately retires seven authored/expanded cases in four whole suites without equivalent replacement. Work-item queries lose real GitHub/GitLab oversized ASCII secret-prefix empty-query and whitespace-before-trim refusal alarms; inputs share the imported size bound and exercise only its fast length branch, with no multibyte or actual provider/exposure guarantee. Skill sharing loses complete/partial managed-bundle and destination-ambiguity association checks plus three negative origin/path/scheme admission examples; these do not establish uploaded contents, arbitrary-URL fetching, authorization or execution. Retain exact originless-project direct-link fallback, cloud-delete confirmation, bundled-target repair/owner isolation and explicit Windows-host preference over stale WSL suites. Rejected3201 runtime-preference removal is excluded. No real cloud/network/host/disk, mobile or mixed-version execution validation is claimed. Audit3200 composes approved source3194: deliberately retires thirteen authored registrations/twenty-seven expanded cases in six whole automation host/list policy suites without equivalent replacement. Real RPC category/retryability and wrapped-timeout recognition, unscoped-failure/scoped-success health evidence precedence, exact-ID create selectors and per-run workspace omission, paired-web local/SSH/absent host catalog routing, incompatible/removed/unhydrated/offline execution health, runtime-before-nested-SSH recovery dispatch, and no-owned-focus/successor/foreign-list focus policy alarms are consciously lost. Retain exact20line orphan allow/refuse/precondition and exact90line cross-host same-ID search suites. No retry-timer/fence lifecycle, actual reconnect/network/write delivery, native focus, mobile/mixed-version runtime or disk persistence coverage is claimed. Audit3191 combines approved source3182/3187/3188: deliberately retires thirteen authored Vitest registrations and fifteen expanded cases, including five whole suites, without equivalent replacement. Attachment intake loses real portal/native-capture terminal interception and unowned-container refusal, immediate child-only committed availability/claim-ref negotiation and delivery, and direct-target over stable focused/background composer routing through disable/re-enable/remount and independent question focus. Exact destination/cache/draft and final-stat owner/workspace/File-release suites remain; they do not replace Surface recipient routing. Automation authoring loses destination runtime catalog membership, local/direct-SSH exclusion of paired-runtime repos, unknown-runtime empty catalog, logical project/host/setup/repo/path context payload, and missing/nonready setup refusal; these were genuine direct helper alarms, while exact schedule transformation, legacy schedule and setup-decision suites remain. Agent picker loses setup-chat/terminal/missing-installer/unsupported-or-unknown-host row projection samples and the ordinary unsearched menu variant; the entire original26-line resume recipe suite remains, including reported bypass versus default/absent permission modes, as does the exact search-result menu callback with its fixtures, model/effort commits and staged Arguments reopen outcome. Rejected predecessor3180 full recipe deletion is excluded. No actual native drag/cursor/focus/IME, upload bytes or disk/restart persistence, full automation save/scheduler/runtime transport, real resume-process flags/environment delivery, or native menu/rendered layout validation is claimed. Audit3178 combines source3169/3170/3175: deliberately retires eight authored units (six Vitest cases and two imperative custom relay scenarios), including two whole suites, without equivalent replacement. Voice Section loses per-model pending-delete isolation across interleaved A/B completion and the two-success refresh-call alarm; exact VoicePane delayed key-clear/newer enabled-toggle merge remains and does not replace row deletion coverage. Speech loses dedicated rating-scale/inclusive-ceiling, unsupported-versus-unmeasured, explicit/system-language precedence/locale alias, accuracy versus fast named/default/best-measured selection, and unmeasured/no-match examples; exact offered-user-pick versus retired-selection case and its fixture remain. Relay loses legacy ping-delivery count/liveness sampled before the silence deadline and grouped malformed Bearer/host-ID 401/429 routing checks; twelve real Worker/DurableObject/shipped-client scenarios remain, not equivalent replacement for those narrow alarms. Six prior speech cases are not all deleted: approved successor3175 keeps the original user-choice case. No actual credential/file persistence, native/GPU/engine/download delivery, production assignment policy or deployed-runtime validation is claimed by this reduction. Audit3165 combines source3160/3162: deliberately retires six authored/expanded cases across three whole suites without equivalent replacement. Launch/status losses: observed React-render pending/resuming coordination across settings and deferred offer read; stale-offer reread suppressing dialog reopen; held-local-chat recovery visibility with new-chat preferences off; disabled-local admission avoiding offer RPC despite an injected but unconsumed paired-tab fixture; and run-only lost-reply uncertainty reachability until direct viewed-run retirement. Launch probe does not prove arbitrary synchronous-store atomicity, the paired fixture is not remote runtime enrollment, and the status case directly marks viewed state rather than rendering the summary. Crew-hook losses: old-host method_not_found cache persists across party changes but is released by independently supplied connection generation/verified pairing status or pairing revision, with actual frame publication and teardown. These are real narrow coordination, admission, uncertainty and mixed-version cache alarms consciously lost. Exact web verified-runtime writer handoff guard remains; it does not replace those alarms or prove live host/native/mobile delivery. Audit3156 combines3147/3150/3152/3153/3154: deliberately retires39authored/40expanded native-chat recovery cases, eight whole suites and two orphan test-support files without equivalent replacement. Lost alarms: lost-reply unconfirmed reconciliation, skipped rows/counts and missed-frame repair, unknown/pending outcome and omitted-failure-list compatibility, uncertain history after dismissal, retry/refusal/dismiss lifecycle, hidden/consumed picker ownership, result-tree keyboard focus/disclosure and all-skipped retirement; unconfirmed-over-conflict guidance and failed-chat changes despite older host timestamps; nearest-listed ancestor/cyclic forest shape, descendant membership, host ordering/missing-host local fallback; skipped/readiness/admission counts, provisional skip/progress monotonicity, early per-chat verdicts with batch pending protection, older-host waiting, start-slot/delivery distinction, retry-over-failure and unknown-after-withdrawal attention; remote snooze/reconnect/reoffer, unseen mixed summary through open/close, watched completion retirement, new success beside old failure, failed-reread mixed-result preservation, filter/collapse completion resets, retry summary replacement and current-offer/history selection; cross-host workspace identity, rendered cyclic reachability, nested tri-state/toggle propagation, project eligibility isolation and collapse preserving selection. Exact fallback-list completion barrier, batched verdict/cleanup retention, stale-read versus dismissal publication race, and requested-but-never-rendered finished-run retirement remain; they do not replace retired alarms. Audit3146 combines source reviews3132/3133/3135/3138/3143: intentionally retire20authored/31expanded cases without equivalent replacement. Provider-session losses: structured-native terminal-resume exclusion, metadata-less completed identity, new-turn completion clearing, Pi provider-only/periodic/manual promotion, changed-transcript launch-config isolation, Pi/omp/prime-agent completed-session remote/manual-shutdown promotion, and quit/done precedence over heartbeat. Side-thread/question losses: finished-pane acquisition, hidden title/recap states, local/relay/relay-done conversation takeover with prompt/session/child resets, plain-message abandonment, compaction/session replacement and completion resets. Codex startup losses: incoming-owner cached child replacement, old-relay foreign live-session correction, completed foreign-root turn-fact permutations. Transcript losses: real temporary-JSONL partial sibling/final-child completion, unlabelled child attribution, compact cached replay preservation and100/1000completed-child roster retention/release. Compiler-ratchet losses: valid-line uncompiled-list parser grammar examples and injected-existence stale-path filtering; these did not exercise actual filesystem canonicalization or build exclusion. These are real lifecycle, ownership, replay, liveness and retention alarms, including recent#40246 regressions, deliberately lost. Retained conditional record replacement, clockless root/SSH owner isolation, persisted main-thread fork rejection, partial/final question-answer lifetime, cached-roster/missing-ID/live HTTP/durable startup boundaries and restarted-relay refusal/acceptance do not replace those losses. No native/SSH/WSL execution or equivalent coverage is claimed. Audit3128 scoped coverage reduction: Audit3121 removes VoiceDictationLanguageSetting.test.tsx (Spanish-to-es callback and unsupported-model Korean option presence) and voice-speech-model-rows.test.ts (provider/language-order/retired-on-disk list policy); these real narrow alarms are deliberately retired without replacement. VoiceSpeechModelSection per-model concurrent deletion remains unchanged. Audit3122 removes the direct absolute/parent-segment sparse-directory exception-type sample and the five-row IPC sparse-preset error-code/non-save table (two authored/six expanded cases). Those real typed-validation and IPC refusal/error-code alarms are deliberately retired. Keep both real symlink/junction floating-directory trust cases and the minimal RPC structured validation-reason serialization case. The RPC case does not replace IPC save/refusal behavior, typed normalization for both path inputs, actual Git sparse checkout or rendered localization. No equivalent replacement or fresh test/runtime/platform/remote/performance validation is claimed. Audit3124 removes ExperimentalPane.test.tsx (ordinary Pet switch callback), WorkspaceDirectorySetting.test.tsx (padded absolute saved-path normalization/no-write blur), and repository-search.perf.test.ts (navigation/selected-pane builder-call policy, fixture cardinality/nonempty entries and renamed project keyword inclusion). These real narrow alarms are retired without replacement; RepositoryPane.search.test.tsx search-to-sparse-editor request remains unchanged. Audit3129: resolver Windows case-insensitive and POSIX case-sensitive configured-env composition plus mocked-base launch-preflight Windows composition are intentionally no longer checked by the two deleted suites. No equivalent coverage claimed. Keep runtime-structured-agent-host-env.test.ts unchanged: configured Windows Claude-home account identity and host-account precedence remain checked. Audit3114 incorporates source3101: retire three structured Claude child-environment wrapper cases covering inherited stamp isolation/default and explicit entrypoint, Windows SDK-key canonicalization/default/duplicate-key order, and configured lookup/final environment/config-dir conjunction. Keep direct auth/session sanitation and Windows-versus-Linux overlay checks, actual executable lookup from supplied inherited PATH, native Windows child environment delivery, and stale option-mutation fencing. These are not equivalent replacements for the retired wrapper alarms. No fresh test/runtime, real login-shell probe, Claude SDK startup, remote/mixed-version, UI or performance result is claimed. Audit3114 incorporates source3105: retire 17 authored/20 expanded daemon launch-note cases, including the entire two-case SessionLaunchNote suite. Deliberately lose split-marker carry across same-token re-arm, stale-token retirement refusal, foreground tri-state and serial settlement, omitted-note/no-incarnation mixed-version policy, sticky retirement, first-incarnation/replacement launch preservation and held-launch precedence. Keep only the unchanged unresolved foreground-read race against a later finished attach. Supplied notes/foreground and runtime calls do not establish actual daemon/main restart, SSH, native Windows observation, mobile UI, equivalent replacement or fresh runtime/performance validation. Audit3114 incorporates source3109: retire six authored/35 expanded pricing cases: 21 accepted model-name examples, 10 unsupported/future-model fail-closed examples, malformed TTL clamp, unknown-model estimate refusal, flat-model long-context disregard and oversized long-context clamp. Keep mixed five-minute/one-hour write allocation and short/all-long/mixed long-context allocation with original fixed oracles. Real parser/refusal/bounds alarms are knowingly lost without equivalent replacement; retained tests do not validate provider rates, actual billing or request threshold classification. Preserve the archived usage-scan-append source-freeze receipt without transferring historical execution/performance credit. No fresh runtime result is claimed. Audit3108 source-only reduction (3094/3095/3100): retire 27 authored/30 expanded cases, including complete markdown view-state, file-preview link-open and CLI agent-server-route suites, without replacement. Deliberate losses include preview visibility/reference lifetime and same-path owners; SSH stat, directory/missing-target, explicit-null owner, nonmarkdown line/column, local URI routing, remote-to-client guard and same-file reuse; document-preview source-owner/folder/reuse routing, stable grant location and outside-workspace POSIX/Windows request projection; and verified remote CLI send/check refusal, suggested --on argv/copy, shell no-probe bypass, unverified-launch admission and older-peer refusal. Retain two markdown source-runtime/qualified-reveal alarms, one stale grant-rejection fence and two remote read/mutation evidence policies, with original callbacks. Final last-wins reveal did not prove transient cancellation; CLI prerequisite did not prove no real send, and envelope tests do not execute server refusal. No fresh runtime, rendered UI, actual file/SSH/paired transport, authorization, mixed-version or performance result is claimed. Audit3096 source-only reduction (3087/3088): retire 69 authored/76 expanded cases, including complete recovery-portals and reconnect-banner suites. Intentionally lose dead/default TUI replay, generated alternate-frame style composition, reset/mode/dimension and legacy missing-mode-boundary alarms, UI notice/action/reveal policies, and isolated recovery lifecycle variants, without equivalent replacement. Retained registrations and callbacks are unchanged; headless cells are not pixels, and no fresh runtime, host, SSH, WSL, rendered-app or performance result is claimed. Audit3086 source-only reduction (3079/3080/3081): intentionally retire 93 authored/100 expanded mock-boundary cases, including the complete spawn-errors and SSH reconnect overlay suites, without replacement. Retired alarms include spawn classification/ownership, reply ceilings and renderer-write failure recovery, disconnect/detach fencing branches, authority facts, and overlay retry/visibility wiring. Retained callbacks are unchanged; no fresh execution, real PTY/SSH, rendered UI, or performance result is claimed. Audit2432: The broad unit command loses the isolated BranchPicker open-change notification sample. Its unique true/false callback alarm is intentionally retired; reopening/status-refresh behavior was not exercised. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2425: The broad unit command loses the workspace-copy-status supplied readError forwarding sample. It did not exercise the branch-switch blocking promised by its title. This alarm is intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2417: The broad unit command loses two ordinary clean Git cherry-pick/revert success samples and two terminal startup foreground-read work-budget samples. These unique alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2412: The broad unit command loses the whole six-case fake-React-hook scroll-anchor suite, two generic draft-retention samples, a simple new-agent order sample, a real history-publication batching-count sample, an unmounted handoff abort/timer sample and three basic WSL preview/error samples. These alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2404: The broad unit command loses the standalone share-name lookup 503 propagation suite, initial deletion placement-kind sample, direct changed-mtime stale sample, first-pass rollback existence example and abandoned-preparation directory sample. These exact alarms are intentionally retired without equivalent replacement or fresh validation of the reduced selector. Audit2397 review amendment: The broad unit command loses three standalone suites, two partial-case definitions and two URL table rows across package/download/placement tests. Keep initial wrong-port authorization, modified independent-copy digest refusal and abort-after-begin cancellation dispatch. The remaining alarms are deliberately retired; no equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2393 current-parent refresh2401: Remove the basic skill-install service alias sample and original public managed-remove receiptless refusal case. Preserve the upstream request ownership suite and transaction interruption coverage. No equivalent complete replacement or fresh validation of the reduced selector is claimed. Audit2389: The broad unit command loses the missing-preview suite, plain-global destination sample and prepublished-receipt recovery example. No equivalent replacement or fresh validation of the reduced selector is claimed. Audit2385: The broad unit command loses four authored walker/staging/protocol examples (six expanded cases); no equivalent replacement or fresh validation is claimed for the reduced suite. Audit2383: The unscoped Vitest command no longer includes the removed Antigravity catalogue suite; its exact provider-root/owner example has no equivalent replacement or fresh validation claim. Audit2377/2381: The unscoped Vitest command includes the reduced bundled-skill and native-discovery suites; the eight removed backup/legacy/catalogue/symlink/SSH-builder examples have no replacement-equivalence or fresh validation claim here. Audit2133 current-parent preparation: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Audit2095 current-parent preparation2126: Current reduced sources/selectors intentionally retire the alarms listed in knownGaps; prior dated runs and replacement claims do not validate the changed set. Historical prior account (preserved verbatim): Require both existing named browser placement scenarios to pass three times with one attempt, zero skips, zero failures, and no report errors.", "commands": [ "gh workflow run packaged-browser-e2e.yml", "pnpm exec playwright test tests/e2e/packaged-mixed-version-browser-placement.spec.ts --config tests/playwright.config.ts --project=electron-headless --workers=1 --repeat-each=3 --retries=0", @@ -14525,7 +14511,8 @@ "Audit 3283 removes 20 authored declarations and 31 expanded scenarios, including two whole suites. Retired alarms cover PR refresh-state pruning, shared runtime subscription cleanup and store isolation, language content-change routing and older completion ownership, ordinary panel retry-timer cleanup and pending loading, four hosted-review lifetime permutations, mocked terminal creation error/order wiring, five sleep-inventory variants with pending-frame-only assertions, palette fallback/composite highlights, Windows/UNC encoded-editor input rows, mixed-repository partial failure counts, title setters and generated-title bypass/replacement, five window-close coordinator delegation/guard/finally branches, and ambiguous-old-record frame reveal. These regression checks are intentionally lost under the test-value audit; retained cases are not equivalent replacements. Prior coverage notes and receipts remain historical.", "Audit 3292 combines adopted 3285/3287/3289/3293 removals: 40 authored declarations and 47 expanded scenarios, including whole frame-editor-view, ambiguous remote-delete-refusal, unverifiable-host subscription and editor-close cleanup suites. The previously reviewed dock-badge suite was already removed upstream in PR 40895 and receives zero new removal credit. Intentionally retired alarms include background browser selection and last-tab landing; main startup fallback, SSH project cleanup/survivor routing, exact-PR contradiction/saved-push positives, replay/sidecar rollback and wrapped remote error normalization; records/snapshot missing-diff hydration, ambiguous folder-owner delete refusal, focus no-op notifications, nested editor context-only cache check, pending broadcast-first status and ordinary four-confirmation aggregation; preview-source frontmatter pruning, approved local/runtime agent dispatch, non-GitHub legacy PR cache invalidation, stale merged-head transient-error fallback, empty-queue immediate query reply and unverified-to-concrete preflight risk. Additional intentionally retired alarms are injected host subscription/unsubscribe counts, synthetic editor reveal/draft/fallback/scoped close, ambiguous unread refusals and hidden mark/read roundtrip, immediate connect backoff/noError, positive mirrored evidence-clock renewal, supplied VM failure/success projections and git-local/folder-SSH wake permutations. Remaining temporal/ownership cases are narrower; no equivalent replacement or fresh execution is claimed.", "Audit 3301 combines adopted source batches 3296 and 3299: retire 27 authored declarations and 36 expanded scenarios, including six whole suites. Intentionally lost alarms cover failed-verification/recovery/restart counters and mocked hosting/toast; early overlapping pane-read scheduling/count; unreadable removal listing draft guard; bulk-purge remote-qualified and raw-local guard maps; initial-snapshot blank/mode/oversize/printable/fresh-spawn callback and replay examples; fallback mobile file close and ordinary MRU; default queue overload/drain; seeded browser-page reuse and supplied missing-page URL recreation; ambiguous sleep/hibernation owner refusal and untouched bindings/origins; ordinary selection/activity/fallback priority; skip-remote success and retained hints; supplied structured-chat unsent retention and scoped removal draft eviction; caller-chat focus precedence; open/draft PR cache eviction and merged-head fallback result inventories. Retained adverse-ordering and ownership cases are narrower and do not establish equivalent replacements. No fresh runtime, native, physical provider, cross-platform, mobile or mixed-version evidence is claimed.", - "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed." + "Audit 3312 composes adopted sources 3302 and 3308: deliberately retires 33 authored declarations and 36 expanded cases, including the whole backend startup-focus, active-browser IPC-close-routing, remote-browser-creation and agent-terminal-presence suites. Intentionally lost alarms cover distinct six-pane reconnect handles; SSH host label/Connect callback, unhydrated visit cleanup, repo cache aliases and hidden worktree/tab/PTY cleanup; adopted-agent startup focus and viewer navigation/reuse; staged, pageless, source, stale and orphan browser-close routing; web local-materialization refusal, desktop incapable-host fallback, failed/thrown remote creation and paired-web sign-in refusal; BOM/NUL and remote/local custom-editor byte-read examples; immediate folder activation and unread clearing; and batching idle/coalescing, accumulated byte ceiling, leading-edge-disabled, standalone synchronous oversize rejection, validation yield, deferred pending-tail order and both viewport-clear examples. The newly reintroduced agent-terminal-presence suite also loses actual terminal-status membership positive and nonterminal-status callback suppression alarms, including the recent 40916 guard regression; the older different 2133 presence-suite history remains historical and receives zero duplicate removal credit. These real regression alarms are intentionally lost. The unchanged survivors provide narrower coverage, with no equivalent replacement or fresh gate, provider, rendered, remote/mobile, mixed-version, cross-platform, soak or measured-budget validation claimed.", + "Audit 3318 composes adopted sources 3310, 3313 and 3317: intentionally retires 71 authored declarations and 102 expanded cases, including whole terminal-detach agent-retention, seeded batch/single open-files equivalence, shallow context-menu shortcut-badge, mocked active-workspace delete-dispatch, workspace session hydration, SSH target cleanup, editor session hydration and worktree terminal purge suites. Lost alarms include chained transfer alias and immediate close/exit retention; batch/single divergence, survivor/selection/map preservation and input nonmutation across 400 imperative seeds (one registered case); disconnected pending-handle retry, authoritative absences/close-record shortcut/older-host absence and published-presence/unpublished-frame examples; published empty host-graph omission to pane exit; legacy search ranking/exclusion fallback; provider request upgrade/dedupe/finalizer ownership; asynchronous PR suppression/primary-change and same-number different-repository URL rejection. Deletion routing loses host/instance qualification, paired and legacy mapping, folder ownership, active-versus-hover/menu-row selection, ambiguity, modal/editable/creation/view/main/deleting/missing guards and shortcut advertisement. Unchanged real keyboard/hover/confirmation/catalog E2E remains narrower, not an equivalent replacement. Additional losses include scripted old-host disk/UTF-8/noop/baseline and landed/unlanded lost-reply/replacement-retry outcomes; ordinary/repeated healthy-link reattach; cross-owner terminal/chat binding; oversized replay permanent-unavailable classification; floating reveal keyboard ownership/close/disable/listener disposal; naming/retained-turn/Sleep/error/close/reject status tables; SSH membership prune/new-target-only reads and environment/local bucket isolation; same-frame receipt capacity and positive bare/headless publisher readmission. These actual regression alarms are deliberately lost; no fresh gate, provider, rendered, mobile, mixed-version, cross-platform, soak or measured-budget validation is claimed." ], "demotionRule": "Keep experimental if any direction skips or fails; do not extend timeouts or retry to green." }, diff --git a/config/scripts/ci/sharding/ci-shard-timings.json b/config/scripts/ci/sharding/ci-shard-timings.json index c468fdc0660..b583ee3ab62 100644 --- a/config/scripts/ci/sharding/ci-shard-timings.json +++ b/config/scripts/ci/sharding/ci-shard-timings.json @@ -6488,7 +6488,6 @@ "src/renderer/src/runtime/browser/creation/web-runtime-browser-creation-placement.test.ts": 1927, "src/renderer/src/runtime/browser/creation/web-runtime-browser-link-open.test.ts": 2217, "src/renderer/src/runtime/web-session-tabs-sync/closing/web-session-close-replay.test.ts": 1119, - "src/renderer/src/runtime/web-session-tabs-sync/placement/web-session-open-files-batch-equivalence.test.ts": 1478, "src/renderer/src/runtime/web-session-tabs-sync/identity/web-session-restored-placeholder-identity.test.ts": 985, "src/renderer/src/runtime/web-session-tabs-sync/agent-status/web-session-tabs-agent-completion-notifications.test.ts": 1953, "src/renderer/src/runtime/web-session-tabs-sync/tracking/web-session-tabs-removed-frame-retires-live-publisher.test.ts": 854, @@ -6692,14 +6691,11 @@ "src/renderer/src/store/slices/settings/settings-terminal-quick-commands.test.ts": 1734, "src/renderer/src/store/slices/settings/settings.test.ts": 1938, "src/renderer/src/store/slices/worktrees/create/sparse-presets.test.ts": 217, - "src/renderer/src/store/slices/ssh/ssh.test.ts": 1843, "src/renderer/src/lib/workspace/activation/active-worktree-removal-successor.test.ts": 1867, "src/renderer/src/store/terminals/store-create-tab-start-in-main.test.ts": 1801, "src/renderer/src/store/slices/tabs/store-open-to-side-layout-owner.test.ts": 1919, "src/renderer/src/store/repos/store-session-cascades.test.ts": 1775, - "src/renderer/src/store/slices/editor/hydration/store-session-editor-hydration.test.ts": 1719, "src/renderer/src/store/terminals/store-session-terminal-reconnect.test.ts": 2279, - "src/renderer/src/store/terminals/store-session-workspace-hydration.test.ts": 1733, "src/renderer/src/store/terminals/store-sleep-agent-status-retention.test.ts": 7641, "src/renderer/src/store/terminals/store-sleep-exact-runtime-stop.test.ts": 1842, "src/renderer/src/store/terminals/store-sleep-pane-hibernation.test.ts": 1932, @@ -6719,7 +6715,6 @@ "src/renderer/src/store/slices/tabs/tabs-reconciliation-empty-groups.test.ts": 1703, "src/renderer/src/store/terminals/terminal-layout-pty-ownership.test.ts": 1734, "src/renderer/src/store/terminals/terminal-pane-detach-agent-identity.test.ts": 1627, - "src/renderer/src/store/terminals/terminal-pane-detach-agent-retention.test.ts": 2846, "src/renderer/src/store/terminals/terminal-pty-identity-replacement.test.ts": 1751, "src/renderer/src/store/terminals/terminal-quick-command-hosts.test.ts": 1759, "src/renderer/src/store/terminals/terminal-tab-owner-index.test.ts": 50, @@ -6768,7 +6763,6 @@ "src/renderer/src/store/slices/worktrees/metadata/worktrees-linked-review-push-target.test.ts": 1186, "src/renderer/src/store/slices/worktrees/metadata/worktrees-multiple-attachments.test.ts": 1132, "src/renderer/src/store/slices/worktrees/create/worktrees-pending-creation-state.test.ts": 1132, - "src/renderer/src/store/slices/worktrees/teardown/worktrees-purge-terminal-state.test.ts": 1133, "src/renderer/src/store/slices/worktrees/create/worktrees-remote-runtime-create.test.ts": 1244, "src/renderer/src/store/slices/worktrees/teardown/worktrees-remote-runtime-removal.test.ts": 1227, "src/renderer/src/store/slices/worktrees/teardown/worktrees-removal-state-cleanup.test.ts": 1229, diff --git a/src/renderer/src/components/sidebar/workspace-actions/context-menu/WorktreeContextMenuDeleteItems.test.tsx b/src/renderer/src/components/sidebar/workspace-actions/context-menu/WorktreeContextMenuDeleteItems.test.tsx deleted file mode 100644 index 90cbcff2ec2..00000000000 --- a/src/renderer/src/components/sidebar/workspace-actions/context-menu/WorktreeContextMenuDeleteItems.test.tsx +++ /dev/null @@ -1,78 +0,0 @@ -// @vitest-environment happy-dom -import { act, type ReactNode } from 'react' -import { createRoot } from 'react-dom/client' -import { expect, it, vi } from 'vitest' -import type { AppState } from '@/store/types' -import { makeWorktree } from '@/store/slices/worktrees/worktree.test-fixture' -import type { WorktreeContextMenuModel } from './use-worktree-context-menu-model' - -// oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: replaced before rendering with every store field read by the production selector. -const mocks = vi.hoisted(() => ({ state: {} as AppState })) -vi.mock('@/store', () => ({ - useAppStore: (select: (s: AppState) => unknown) => select(mocks.state) -})) -vi.mock('@/hooks/useShortcutLabel', () => ({ useOptionalShortcutLabel: () => '⌘⇧⌫' })) -vi.mock('@/i18n/use-translate', () => ({ - useTranslate: () => (_key: string, fallback: string) => fallback -})) -vi.mock('@/components/ui/dropdown-menu', () => ({ - DropdownMenuItem: ({ children }: { children: ReactNode }) =>
{children}
, - DropdownMenuShortcut: ({ children }: { children: ReactNode }) => {children} -})) -import { WorktreeContextMenuDeleteItems } from './WorktreeContextMenuDeleteItems' - -it('advertises the shortcut only on the committed paired copy when physical host and instance match', () => { - const inactive = makeWorktree({ - id: 'repo::/checkout', - repoId: 'repo', - hostId: 'ssh:builder', - runtimeOwnerEnvironmentId: 'server-b' - }) - const active = { ...inactive, runtimeOwnerEnvironmentId: 'server-a' } - // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: the isolated store mock supplies every field read by the production selector. - mocks.state = { - activeView: 'terminal', - activeWorktreeId: active.id, - activeWorkspaceExecutionHostId: 'runtime:server-a', - activePendingCreationId: null, - pendingWorktreeCreations: {}, - worktreesByRepo: { repo: [inactive, active] }, - repos: [], - folderWorkspaces: [], - projectGroups: [], - deleteStateByWorktreeId: {} - } as unknown as AppState - const container = document.createElement('div') - document.body.append(container) - const root = createRoot(container) - const model = (worktree: typeof inactive) => - // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: this fixture supplies every model field read by DeleteItems; other menu sections never render. - ({ - worktree, - batchDeleteWorktrees: [], - contextDeletePending: false, - handleDelete: vi.fn(), - isMultiContext: false, - lineageDescendantCount: 0, - removesProject: false - }) as unknown as WorktreeContextMenuModel - try { - act(() => - root.render( - <> -
- -
-
- -
- - ) - ) - expect(container.querySelector('[data-copy="inactive"] kbd')).toBeNull() - expect(container.querySelector('[data-copy="active"] kbd')).not.toBeNull() - } finally { - act(() => root.unmount()) - container.remove() - } -}) diff --git a/src/renderer/src/components/sidebar/workspace-actions/deletion/active-workspace-delete.test.ts b/src/renderer/src/components/sidebar/workspace-actions/deletion/active-workspace-delete.test.ts deleted file mode 100644 index 92643804fad..00000000000 --- a/src/renderer/src/components/sidebar/workspace-actions/deletion/active-workspace-delete.test.ts +++ /dev/null @@ -1,283 +0,0 @@ -// @vitest-environment happy-dom - -import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest' -import type { AppState } from '@/store/types' -import { makeWorktree } from '@/store/slices/worktrees/worktree.test-fixture' -import { makeFolderWorkspace } from '@/store/folder-workspaces/folder-workspace.test-fixture' -import { TEST_REPO } from '@/store/repos/repo.test-fixture' - -const mocks = vi.hoisted(() => ({ - // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: replaced before each test with every store field read by these shortcut callers. - state: {} as AppState, - deleteWorktree: vi.fn(), - folder: vi.fn() -})) -vi.mock('@/store', () => ({ useAppStore: { getState: () => mocks.state } })) -vi.mock('./delete-worktree-flow', () => ({ runWorktreeDelete: mocks.deleteWorktree })) -vi.mock('./folder-workspace-delete', () => ({ deleteFolderWorkspaceUnlessUnsaved: mocks.folder })) -vi.mock('@/components/floating-terminal/floating-workspace-terminal-actions', () => ({ - isFloatingWorkspacePanelFocused: () => false -})) -import { createAppCommandHandlers, type AppShortcutState } from '@/app-shell/app-command-handlers' -import { registerWorkspaceShortcutIpcBridge } from '@/hooks/ipc-events/workspace-shortcut-ipc-bridge' -import { - createWorktreeContextMenuDeleteIntent, - runWorktreeContextMenuDeleteIntent -} from '../context-menu/worktree-context-menu-delete-intent' - -const active = makeWorktree({ id: 'active', repoId: 'repo1', hostId: 'local', instanceId: 'a' }) -const other = makeWorktree({ id: 'other', repoId: 'repo1', hostId: 'local', instanceId: 'b' }) -let ipcDelete: () => void - -function runRenderer(): void { - const state: AppShortcutState = { - activeView: 'terminal', - activeWorktreeId: 'active', - actions: { - toggleSidebar: vi.fn(), - toggleRightSidebar: vi.fn(), - setRightSidebarOpen: vi.fn(), - setRightSidebarTab: vi.fn(), - showRightSidebarFiles: vi.fn(), - showRightSidebarSearch: vi.fn(), - openDiffNotesSendMenuForActiveWorktree: vi.fn() - }, - creationLayoutActive: false, - floatingTerminalEnabled: false, - floatingTerminalOpen: false, - floatingVisibleTabCount: 0, - keybindings: {}, - openFloatingWorkspaceMaximized: vi.fn(), - pluginCommands: [], - setFloatingTerminalOpen: vi.fn(), - terminalShortcutPolicy: 'orca-first', - workspaceChromeActive: true - } - createAppCommandHandlers(state, undefined, 'app', true).get('workspace.delete')?.() -} - -beforeEach(() => { - vi.clearAllMocks() - mocks.folder.mockReturnValue(null) - // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: this isolated store mock supplies every field read by the tested production paths. - mocks.state = { - activeView: 'terminal', - activeModal: 'none', - activeWorktreeId: 'active', - activeWorkspaceExecutionHostId: 'local', - activePendingCreationId: null, - pendingWorktreeCreations: {}, - repos: [TEST_REPO], - worktreesByRepo: { repo1: [active, other] }, - folderWorkspaces: [], - projectGroups: [], - deleteStateByWorktreeId: {} - } as unknown as AppState - const noop = () => () => {} - Object.defineProperty(window, 'api', { - configurable: true, - value: { - ui: { - onOpenQuickOpen: noop, - onToggleQuickCommandsMenu: noop, - onOpenNewWorkspace: noop, - onDeleteCurrentWorkspace: (callback: () => void) => { - ipcDelete = callback - return () => {} - }, - onOpenTasks: noop, - onJumpToWorktreeIndex: noop, - onJumpToTabIndex: noop, - onWorktreeHistoryNavigate: noop, - onToggleStatusBar: noop - } - } - }) - registerWorkspaceShortcutIpcBridge([]) - const query = document.querySelectorAll.bind(document) - vi.spyOn(document, 'querySelectorAll').mockImplementation((selector) => - query(selector.replace(':hover', '[data-hovered]')) - ) - document.body.innerHTML = - '
' -}) -afterEach(() => { - document.body.innerHTML = '' - vi.restoreAllMocks() -}) - -describe.each([ - ['renderer', runRenderer], - ['main IPC', () => ipcDelete()] -] as const)('%s delete shortcut', (_name, run) => { - it('targets the active instance while a different row is hovered, but menu clicks keep their row', () => { - run() - expect(mocks.deleteWorktree).toHaveBeenCalledExactlyOnceWith('active', { - expectedHostId: 'local', - expectedInstanceId: 'a' - }) - runWorktreeContextMenuDeleteIntent( - createWorktreeContextMenuDeleteIntent({ - worktree: other, - batchDeleteWorktrees: [], - isMultiContext: false - }) - ) - expect(mocks.deleteWorktree).toHaveBeenLastCalledWith('other', { - expectedHostId: 'local', - expectedInstanceId: 'b' - }) - }) - - it('works with no sidebar or hovered row', () => { - document.body.innerHTML = '' - run() - expect(mocks.deleteWorktree).toHaveBeenCalledExactlyOnceWith('active', { - expectedHostId: 'local', - expectedInstanceId: 'a' - }) - }) - - it('pins the active host when another host publishes the same ID', () => { - mocks.state.worktreesByRepo = { - repo1: [active, { ...active, hostId: 'ssh:builder', instanceId: 'remote' }] - } - mocks.state.activeWorkspaceExecutionHostId = 'ssh:builder' - run() - expect(mocks.deleteWorktree).toHaveBeenCalledExactlyOnceWith('active', { - expectedHostId: 'ssh:builder', - expectedInstanceId: 'remote' - }) - }) - - it('resolves legacy local ownership without inventing a stamped host', () => { - mocks.state.worktreesByRepo = { repo1: [{ ...active, hostId: undefined }] } - run() - expect(mocks.deleteWorktree).toHaveBeenCalledExactlyOnceWith('active', { - expectedInstanceId: 'a' - }) - }) - - it('keeps a paired logical host bound to its physical row', () => { - mocks.state.worktreesByRepo = { - repo1: [{ ...active, hostId: 'ssh:builder', runtimeOwnerEnvironmentId: 'build' }] - } - mocks.state.activeWorkspaceExecutionHostId = 'runtime:build' - run() - expect(mocks.deleteWorktree).toHaveBeenCalledExactlyOnceWith('active', { - expectedHostId: 'ssh:builder', - expectedInstanceId: 'a' - }) - }) - - it('refuses an ambiguous unqualified active ID', () => { - mocks.state.activeWorkspaceExecutionHostId = null - mocks.state.worktreesByRepo = { repo1: [active, { ...active, hostId: 'ssh:builder' }] } - run() - expect(mocks.deleteWorktree).not.toHaveBeenCalled() - }) - - it('routes the active folder to its qualified owner', () => { - mocks.state.activeWorktreeId = 'folder:folder-1' - mocks.state.activeWorkspaceExecutionHostId = 'ssh:builder' - mocks.state.folderWorkspaces = [ - makeFolderWorkspace({ executionHostId: 'local' }), - makeFolderWorkspace({ executionHostId: 'ssh:builder' }) - ] - run() - expect(mocks.folder).toHaveBeenCalledExactlyOnceWith( - expect.any(Function), - 'folder-1', - 'ssh:builder' - ) - expect(mocks.deleteWorktree).not.toHaveBeenCalled() - }) - - it.each(['settings', 'activity', 'space', 'tasks'] as const)( - 'refuses behind the %s view', - (activeView) => { - mocks.state.activeView = activeView - run() - expect(mocks.deleteWorktree).not.toHaveBeenCalled() - } - ) - - it('refuses while a creation surface owns the workspace', () => { - mocks.state.activePendingCreationId = 'pending' - mocks.state.pendingWorktreeCreations = { - // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: creation eligibility reads only status and loaderVisible. - pending: { - status: 'error', - loaderVisible: true - } as AppState['pendingWorktreeCreations'][string] - } - run() - expect(mocks.deleteWorktree).not.toHaveBeenCalled() - }) - - it('refuses a local dialog with button focus even when activeModal is none', () => { - const dialog = document.createElement('div') - dialog.role = 'dialog' - dialog.innerHTML = '' - document.body.append(dialog) - // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: overlay visibility reads only the returned rectangle count. - vi.spyOn(dialog, 'getClientRects').mockReturnValue([ - new DOMRect(0, 0, 300, 200) - ] as unknown as DOMRectList) - dialog.querySelector('button')?.focus() - run() - expect(mocks.deleteWorktree).not.toHaveBeenCalled() - }) - - it('allows an ordinary context menu without changing the active target', () => { - document.body.insertAdjacentHTML( - 'beforeend', - '
' - ) - const menu = document.querySelector('[role="menu"]') - if (!menu) { - throw new Error('Missing menu') - } - // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: overlay visibility reads only the returned rectangle count. - vi.spyOn(menu, 'getClientRects').mockReturnValue([ - new DOMRect(0, 0, 300, 200) - ] as unknown as DOMRectList) - document.querySelector('button')?.focus() - run() - expect(mocks.deleteWorktree).toHaveBeenCalledExactlyOnceWith('active', { - expectedHostId: 'local', - expectedInstanceId: 'a' - }) - }) - - it.each(['missing', 'primary', 'deleting', 'modal', 'editable'] as const)( - 'refuses %s targets or focus', - (guard) => { - if (guard === 'missing') { - mocks.state.activeWorktreeId = 'missing' - } - if (guard === 'primary') { - mocks.state.worktreesByRepo = { repo1: [{ ...active, isMainWorktree: true }] } - } - if (guard === 'deleting') { - mocks.state.deleteStateByWorktreeId = { - 'local|active': { - isDeleting: true, - error: null, - canForceDelete: false, - forceDeleteReason: null - } - } - } - if (guard === 'modal') { - mocks.state.activeModal = 'delete-worktree' - } - if (guard === 'editable') { - document.body.insertAdjacentHTML('beforeend', '') - document.querySelector('input')?.focus() - } - run() - expect(mocks.deleteWorktree).not.toHaveBeenCalled() - } - ) -}) diff --git a/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-healthy-link-reattach.test.ts b/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-healthy-link-reattach.test.ts index 7aa82476efb..2190efb447b 100644 --- a/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-healthy-link-reattach.test.ts +++ b/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-healthy-link-reattach.test.ts @@ -99,62 +99,6 @@ describe('remote runtime PTY reattach on a healthy link', () => { return transport } - it('reattaches a handle the host still reports ready within seconds of a stream end', async () => { - const transport = await attachHostMirror() - runtimeCall.mockImplementation(async (args: { method: string }) => - args.method === 'session.tabs.list' || args.method === 'session.tabs.activate' - ? readyHostSessionInventoryResponse('terminal-1') - : { ok: true, result: {} } - ) - - endLatestStream() - await vi.advanceTimersByTimeAsync(2_000) - - expect(subscribedTerminalHandles()).toEqual(['terminal-1', 'terminal-1']) - emitSnapshot(latestSubscribePayload().streamId, 'reattached') - expect(transport.isConnected()).toBe(true) - expect(transport.getRecoveryState?.().phase).toBe('connected') - transport.destroy?.() - }) - - it('keeps reattaching a still-ready handle that ends repeatedly instead of parking', async () => { - const transport = await attachHostMirror() - runtimeCall.mockImplementation(async (args: { method: string }) => - args.method === 'session.tabs.list' || args.method === 'session.tabs.activate' - ? readyHostSessionInventoryResponse('terminal-1') - : { ok: true, result: {} } - ) - - for (let cycle = 0; cycle < 5; cycle += 1) { - endLatestStream() - await vi.advanceTimersByTimeAsync(10_000) - expect(subscribedTerminalHandles()).toHaveLength(cycle + 2) - emitSnapshot(latestSubscribePayload().streamId, `cycle ${cycle}`) - expect(transport.isConnected()).toBe(true) - expect(transport.getRecoveryState?.().phase).toBe('connected') - } - transport.destroy?.() - }) - - it('retries a resubscribe whose host answered without a ready handle', async () => { - const transport = await attachHostMirror() - let hostReady = false - runtimeCall.mockImplementation(async (args: { method: string }) => - args.method === 'session.tabs.list' || args.method === 'session.tabs.activate' - ? hostSnapshot(hostReady ? 'ready' : 'pending-handle') - : { ok: true, result: {} } - ) - - endLatestStream() - await vi.advanceTimersByTimeAsync(16_000) - expect(subscribedTerminalHandles()).toHaveLength(1) - hostReady = true - await vi.advanceTimersByTimeAsync(5_000) - - expect(subscribedTerminalHandles()).toEqual(['terminal-1', 'terminal-1']) - transport.destroy?.() - }) - it('retries an attach whose host surface stayed pending past the bounded wait', async () => { let hostReady = false runtimeCall.mockImplementation(async (args: { method: string }) => diff --git a/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-unpublished-host-graph.test.ts b/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-unpublished-host-graph.test.ts index edfb657bc89..5f2e36380cc 100644 --- a/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-unpublished-host-graph.test.ts +++ b/src/renderer/src/components/terminal-pane/pty-transport/remote-runtime/remote-runtime-pty-transport-unpublished-host-graph.test.ts @@ -175,34 +175,4 @@ describe('remote runtime pty transport against a relaunched host that has not pu vi.useRealTimers() } }) - - it('still retires the pane when a published graph lacks its surface', async () => { - const { createRemoteRuntimePtyTransport } = await import('./remote-runtime-pty-transport') - const onPtyExit = vi.fn() - const transport = createRemoteRuntimePtyTransport('env-1', { - worktreeId: 'wt-1', - tabId: 'web-terminal-host-tab-1', - leafId: 'pane:1', - onPtyExit - }) - await transport.connect({ url: '', callbacks: {} }) - await vi.waitFor(() => expect(subscriptionSendBinary).toHaveBeenCalled()) - runtimeCall.mockImplementation(async (request: { method: string }) => - request.method === 'session.tabs.list' - ? { - ok: true, - result: { - ...UNPUBLISHED_HOST_GRAPH.result, - publicationEpoch: 'epoch-2:client-navigation', - snapshotVersion: 3 - } - } - : { ok: false, error: { code: 'runtime_error', message: 'tab_not_found' } } - ) - - subscriptionCallbacks?.onClose?.() - - await vi.waitFor(() => expect(onPtyExit).toHaveBeenCalledWith('remote:env-1@@terminal-1', -1)) - expect(transport.getPtyId()).toBeNull() - }) }) diff --git a/src/renderer/src/hooks/ipc-events/runtime-file-position.test.ts b/src/renderer/src/hooks/ipc-events/runtime-file-position.test.ts index db849dfe897..12c8f4b80b8 100644 --- a/src/renderer/src/hooks/ipc-events/runtime-file-position.test.ts +++ b/src/renderer/src/hooks/ipc-events/runtime-file-position.test.ts @@ -92,39 +92,6 @@ function focus(surface: 'main' | 'floating'): void { document.getElementById(surface)?.focus() } -it('waits for an asynchronously revealed floating panel', async () => { - const { store, fileId, position } = await setup(false) - position(fileId, { line: 300, column: 7 }, true) - flushFrames() - expect(store.getState().pendingEditorReveal).toBeNull() - store.setState({ floatingWorkspacePanelOpen: true }) - flushFrames() - expect(store.getState().pendingEditorReveal).toMatchObject({ fileId, line: 300, column: 7 }) -}) - -it('keeps a background floating position deferred until the panel owns keyboard input', async () => { - const { store, fileId, position } = await setup() - focus('main') - position(fileId, { line: 17 }, false) - flushFrames() - expect(store.getState().pendingEditorReveal).toBeNull() - focus('floating') - flushFrames() - expect(store.getState().pendingEditorReveal).toMatchObject({ fileId, line: 17 }) -}) - -it('does not publish after keyboard input moves from the panel to the main surface', async () => { - const { store, fileId, position } = await setup() - focus('floating') - position(fileId, { line: 17 }, true) - focus('main') - flushFrames() - expect(store.getState().pendingEditorReveal).toBeNull() - focus('floating') - flushFrames() - expect(store.getState().pendingEditorReveal).toMatchObject({ fileId, line: 17 }) -}) - it('cancels a published reveal after an outside click even if DOM focus does not change', async () => { const { store, fileId, position } = await setup() focus('main') @@ -146,35 +113,6 @@ it('does not reclaim input when an asynchronously revealed panel opens after an expect(store.getState().pendingEditorReveal).toBeNull() }) -it('invalidates a published floating reveal when the panel closes or is disabled', async () => { - const { store, fileId, position } = await setup() - position(fileId, { line: 17 }, true) - flushFrames() - const reveal = store.getState().pendingEditorReveal - expect(reveal?.canApply?.()).toBe(true) - store.setState({ floatingWorkspacePanelOpen: false }) - expect(reveal?.canApply?.()).toBe(false) - store.setState({ floatingWorkspacePanelOpen: true }) - const settings = store.getState().settings - if (!settings) { - throw new Error('Settings missing') - } - store.setState({ settings: { ...settings, floatingTerminalEnabled: false } }) - expect(reveal?.canApply?.()).toBe(false) -}) - -it('keeps floating input ownership independent of an unrelated main workspace selection', async () => { - const { store, fileId, position } = await setup() - focus('floating') - position(fileId, { line: 300, column: 7 }, true) - flushFrames() - const reveal = store.getState().pendingEditorReveal - expect(reveal?.canApply?.()).toBe(true) - store.setState({ activeWorktreeId: 'other-main-workspace' }) - expect(document.activeElement?.id).toBe('floating') - expect(reveal?.canApply?.()).toBe(true) -}) - it('scopes the floating reveal to its selected split pane', async () => { const { store, fileId, position } = await setup() const tab = store.getState().getActiveTab(FLOATING) @@ -199,15 +137,3 @@ it('scopes the floating reveal to its selected split pane', async () => { store.getState().focusGroup(FLOATING, siblingGroupId) expect(reveal?.canApply?.()).toBe(false) }) - -it('disposes floating DOM listeners and invalidates an already published reveal', async () => { - const { store, fileId, position } = await setup() - position(fileId, { line: 17 }, true) - flushFrames() - const reveal = store.getState().pendingEditorReveal - unsubs.forEach((unsubscribe) => unsubscribe()) - expect(reveal?.canApply?.()).toBe(false) - focus('floating') - flushFrames() - expect(store.getState().pendingEditorReveal).toBe(reveal) -}) diff --git a/src/renderer/src/hooks/useEditorExternalWatch-subscriptions.test.tsx b/src/renderer/src/hooks/useEditorExternalWatch-subscriptions.test.tsx index 079676b265a..3fb334d1810 100644 --- a/src/renderer/src/hooks/useEditorExternalWatch-subscriptions.test.tsx +++ b/src/renderer/src/hooks/useEditorExternalWatch-subscriptions.test.tsx @@ -122,53 +122,6 @@ describe('useEditorExternalWatch subscriptions', () => { ;(window as unknown as { api?: unknown }).api = previousApi }) - it('unsubscribes an SSH watch and the shared event listener exactly once on unmount', async () => { - subscriptionState.snapshot = { - targets: [ - { - worktreeId: 'wt-ssh', - worktreePath: '/remote/repo', - connectionId: 'ssh-1', - runtimeEnvironmentId: null - } - ], - targetsKey: 'ssh-watch' - } - await act(async () => root.render(createElement(WatchProbe))) - - expect(watchWorktree).toHaveBeenCalledWith({ - worktreePath: '/remote/repo', - connectionId: 'ssh-1' - }) - await act(async () => root.unmount()) - - expect(unwatchWorktree).toHaveBeenCalledTimes(1) - expect(unwatchWorktree).toHaveBeenCalledWith({ - worktreePath: '/remote/repo', - connectionId: 'ssh-1' - }) - expect(unsubscribeFsEvents).toHaveBeenCalledTimes(1) - expect(subscriptionState.disposeEventHandler).toHaveBeenCalledTimes(1) - }) - - it('disposes a runtime subscription that resolves after unmount', async () => { - const pending = deferredRuntimeSubscription() - const unsubscribeRuntime = vi.fn() - subscriptionState.subscribeRuntimeFileChanges.mockReturnValueOnce(pending.promise) - subscriptionState.snapshot = { - targets: [runtimeTarget()], - targetsKey: 'runtime-watch' - } - await act(async () => root.render(createElement(WatchProbe))) - await act(async () => root.unmount()) - - pending.resolve(unsubscribeRuntime) - await act(async () => pending.promise) - - expect(unsubscribeRuntime).toHaveBeenCalledTimes(1) - expect(unwatchWorktree).not.toHaveBeenCalled() - }) - it('cannot let an old runtime subscribe resolution replace a re-added watch', async () => { const stalePending = deferredRuntimeSubscription() const currentPending = deferredRuntimeSubscription() diff --git a/src/renderer/src/runtime/files/mutations/runtime-file-write-disk-precondition.test.ts b/src/renderer/src/runtime/files/mutations/runtime-file-write-disk-precondition.test.ts index 04602942916..2b11191c7d5 100644 --- a/src/renderer/src/runtime/files/mutations/runtime-file-write-disk-precondition.test.ts +++ b/src/renderer/src/runtime/files/mutations/runtime-file-write-disk-precondition.test.ts @@ -88,46 +88,6 @@ function writeCalls(): RuntimeRpcRequest[] { } describe('writeRuntimeFile disk precondition', () => { - it('refuses without writing when an older host holds newer disk content', async () => { - serveHost({ enforcesPrecondition: false, diskContent: 'agent rewrite' }) - - const result = await writeRuntimeFile( - remoteContext, - '/remote/repo/file.ts', - 'edited', - undefined, - precondition - ) - - expect(result).toEqual({ ok: false, diskConflict: 'changed' }) - expect(writeCalls()).toHaveLength(0) - }) - - it.each(['original', 'edited'])( - 'skips identical content on a v1 guarded host for an automatic save (baseline %s)', - async (baseline) => { - serveHost({ enforcesPrecondition: true, identicalNoop: false, diskContent: 'edited' }) - expect( - await writeRuntimeFile(remoteContext, '/remote/repo/file.ts', 'edited', undefined, { - expectedSignature: getDiskBaselineSignature(baseline), - requireLosslessUtf8: true - }) - ).toEqual({ ok: true, unchanged: true }) - expect(writeCalls()).toHaveLength(0) - } - ) - - it('refuses an ambiguous automatic baseline before sending to a v1 host', async () => { - serveHost({ enforcesPrecondition: true, identicalNoop: false, diskContent: 'a�' }) - expect( - await writeRuntimeFile(remoteContext, '/remote/repo/file.ts', 'a�', undefined, { - expectedSignature: getDiskBaselineSignature('a�'), - requireLosslessUtf8: true - }) - ).toEqual({ ok: false, diskConflict: 'not-utf8' }) - expect(writeCalls()).toHaveLength(0) - }) - it('refuses to send when the runtime connection changes after the probe', async () => { serveHost({ enforcesPrecondition: false, @@ -141,26 +101,6 @@ describe('writeRuntimeFile disk precondition', () => { expect(writeCalls()).toHaveLength(0) }) - it('keeps the save when the guarded write landed but its reply timed out', async () => { - serveHost({ enforcesPrecondition: true, diskContent: 'edited' }) - runtimeEnvironmentCall.mockResolvedValue({ - id: 'rpc', - ok: false, - error: { code: 'runtime_timeout', message: 'Timed out waiting for the remote Orca runtime' }, - _meta: { runtimeId: 'remote-runtime' } - }) - - const result = await writeRuntimeFile( - remoteContext, - '/remote/repo/file.ts', - 'edited', - undefined, - precondition - ) - - expect(result).toEqual({ ok: true }) - }) - it('does not take a re-paired host holding the text as proof a lost write landed', async () => { replaceRuntimeEnvironmentRevisions([{ id: 'env-1', createdAt: 1, pairingRevision: 1 }]) serveHost({ enforcesPrecondition: true, diskContent: 'edited' }) @@ -182,43 +122,4 @@ describe('writeRuntimeFile disk precondition', () => { writeRuntimeFile(remoteContext, '/remote/repo/file.ts', 'edited', undefined, precondition) ).rejects.toThrow('Timed out') }) - - it('still fails a timed-out guarded write whose content is not on disk', async () => { - serveHost({ enforcesPrecondition: true, diskContent: 'original' }) - runtimeEnvironmentCall.mockResolvedValue({ - id: 'rpc', - ok: false, - error: { code: 'runtime_timeout', message: 'Timed out waiting for the remote Orca runtime' }, - _meta: { runtimeId: 'remote-runtime' } - }) - - await expect( - writeRuntimeFile(remoteContext, '/remote/repo/file.ts', 'edited', undefined, precondition) - ).rejects.toThrow('Timed out') - }) - - it('recovers when the transport refused to send to a replacement runtime', async () => { - serveHost({ enforcesPrecondition: true }) - runtimeEnvironmentCall - .mockResolvedValueOnce({ - id: 'files.writeIfUnchanged', - ok: false, - error: { - code: 'runtime_environment_changed', - message: 'Runtime environment identity changed; refresh and try again' - }, - _meta: { runtimeId: 'remote-runtime' } - }) - .mockResolvedValue(rpcOk({ ok: true })) - - const result = await writeRuntimeFile( - remoteContext, - '/remote/repo/file.ts', - 'edited', - undefined, - precondition - ) - - expect(result).toEqual({ ok: true }) - }) }) diff --git a/src/renderer/src/runtime/files/search/runtime-file-client-search-listing.test.ts b/src/renderer/src/runtime/files/search/runtime-file-client-search-listing.test.ts index 44325099070..faf5e09326a 100644 --- a/src/renderer/src/runtime/files/search/runtime-file-client-search-listing.test.ts +++ b/src/renderer/src/runtime/files/search/runtime-file-client-search-listing.test.ts @@ -5,116 +5,10 @@ import { runtimeEnvironmentSubscribe, installRuntimeFileClientEnvironment } from '../runtime-file-client-test-harness' -import { replaceRuntimeEnvironmentRevisions } from '../../rpc/environment-revisions/runtime-environment-revision' installRuntimeFileClientEnvironment() describe('runtime file client', () => { - it('falls back to one cached legacy inventory and ranks evolving queries locally', async () => { - replaceRuntimeEnvironmentRevisions([{ id: 'env-1', createdAt: 1 }]) - runtimeEnvironmentCall.mockImplementation(({ method }) => { - if (method === 'files.searchPaths') { - return Promise.resolve({ - id: 'rpc-search-legacy', - ok: false, - error: { code: 'method_not_found', message: 'Unknown method: files.searchPaths' }, - _meta: { runtimeId: 'legacy-runtime' } - }) - } - return Promise.resolve({ - id: 'rpc-list-legacy', - ok: true, - result: { - worktree: 'wt-1', - rootPath: '/remote/repo', - files: [ - { relativePath: 'src/target.ts', basename: 'target.ts', kind: 'text' }, - { relativePath: 'src/other.ts', basename: 'other.ts', kind: 'text' }, - { relativePath: 'nested/target.ts', basename: 'target.ts', kind: 'text' } - ], - totalCount: 7_000, - truncated: true - }, - _meta: { runtimeId: 'legacy-runtime' } - }) - }) - - await expect( - searchRuntimeFilePaths( - { - target: { kind: 'environment' as const, environmentId: 'env-1' }, - worktreeId: 'wt-1', - worktreePath: '/remote/repo' - }, - { - query: 'target', - limit: 32, - excludePaths: ['/remote/repo/nested'] - } - ) - ).resolves.toEqual({ files: ['src/target.ts'], truncated: true }) - - await expect( - searchRuntimeFilePaths( - { - target: { kind: 'environment' as const, environmentId: 'env-1' }, - worktreeId: 'wt-1', - worktreePath: '/remote/repo' - }, - { query: 'other', limit: 32 } - ) - ).resolves.toEqual({ files: ['src/other.ts'], truncated: true }) - }) - - it('re-applies exclusions for legacy runtimes that ignore quick-open fields', async () => { - runtimeEnvironmentCall.mockImplementation(({ method }) => { - if (method === 'files.searchPaths') { - return Promise.resolve({ - id: 'rpc-legacy-search', - ok: true, - result: { - worktree: 'wt-1', - rootPath: '/remote/repo', - files: Array.from({ length: 32 }, (_, index) => ({ - relativePath: `nested/target-${index}.ts`, - basename: `target-${index}.ts`, - kind: 'text' - })), - totalCount: 33, - truncated: true - }, - _meta: { runtimeId: 'legacy-runtime' } - }) - } - return Promise.resolve({ - id: 'rpc-legacy-list', - ok: true, - result: { - worktree: 'wt-1', - rootPath: '/remote/repo', - files: [ - { relativePath: 'src/target.ts', basename: 'target.ts', kind: 'text' }, - { relativePath: 'nested/target.ts', basename: 'target.ts', kind: 'text' } - ], - totalCount: 2, - truncated: false - }, - _meta: { runtimeId: 'legacy-runtime' } - }) - }) - - await expect( - searchRuntimeFilePaths( - { - target: { kind: 'environment' as const, environmentId: 'env-1' }, - worktreeId: 'wt-1', - worktreePath: '/remote/repo' - }, - { query: 'target', limit: 32, excludePaths: ['/remote/repo/nested'] } - ) - ).resolves.toEqual({ files: ['src/target.ts'], truncated: false }) - }) - it('cancels a legacy paired inventory scan when its only search is superseded', async () => { const controller = new AbortController() const unsubscribeSearch = vi.fn() diff --git a/src/renderer/src/runtime/hosted-terminals/recovery/panes/web-session-terminal-pending-handle-recovery.test.ts b/src/renderer/src/runtime/hosted-terminals/recovery/panes/web-session-terminal-pending-handle-recovery.test.ts index 14d365caea7..c77f1fd2cc2 100644 --- a/src/renderer/src/runtime/hosted-terminals/recovery/panes/web-session-terminal-pending-handle-recovery.test.ts +++ b/src/renderer/src/runtime/hosted-terminals/recovery/panes/web-session-terminal-pending-handle-recovery.test.ts @@ -166,97 +166,6 @@ describe('web session pending terminal handle recovery', () => { }) }) - it('retries a disconnected pane resolution for an unchanged replayed snapshot', async () => { - const readySnapshot = { - ...pendingSnapshot(), - publicationEpoch: 'resolved-after-disconnect', - snapshotVersion: 3, - tabs: [ - { - ...pendingSnapshot().tabs[0]!, - status: 'ready' as const, - terminal: TERMINAL_HANDLE - } - ] - } - let resolveAttempts = 0 - const call = vi.fn(async ({ method }: { method: string }) => { - if (method === 'session.tabs.list') { - return { ok: true as const, result: readySnapshot } - } - if (method === 'terminal.resolvePane') { - resolveAttempts += 1 - return { - ok: true as const, - result: { - terminal: { - handle: TERMINAL_HANDLE, - tabId: HOST_TAB_ID, - leafId: LEAF_ID, - ptyId: 'pty-live', - connected: resolveAttempts > 1, - worktreeId: WORKTREE_ID - } - } - } - } - if (method === 'terminal.list') { - return { - ok: true as const, - result: { - terminals: [ - { - handle: TERMINAL_HANDLE, - ptyId: 'pty-live', - incarnationId: 'inc-live', - orphaned: true - } - ], - topologyRevisions: { [WORKTREE_ID]: 4 }, - totalCount: 1, - truncated: false, - hostScope: { hostIds: ['local'], omittedHostIds: [] } - } - } - } - return { - ok: true as const, - result: { adopted: true, topologyRevision: 5, snapshot: readySnapshot } - } - }) - - await expect( - recoverWebSessionTerminalOrphansBeforeApply( - stateWithEmptyBinding(), - pendingSnapshot(), - ENVIRONMENT_ID, - { call: call as never } - ) - ).resolves.toMatchObject({ - tabs: [expect.objectContaining({ status: 'pending-handle', terminal: null })] - }) - await expect( - recoverWebSessionTerminalOrphansBeforeApply( - stateWithEmptyBinding(), - pendingSnapshot(), - ENVIRONMENT_ID, - { call: call as never } - ) - ).resolves.toMatchObject({ - publicationEpoch: 'resolved-after-disconnect', - snapshotVersion: 3, - tabs: [ - { - id: 'host-tab::11111111-1111-4111-8111-111111111111', - leafId: '11111111-1111-4111-8111-111111111111', - ptyId: 'pty-live', - status: 'ready', - terminal: 'term-live' - } - ] - }) - }) - describe('an unbound pane whose host tab was closed (#9194)', () => { const closedPaneCall = vi.fn() beforeEach(() => { @@ -284,68 +193,10 @@ describe('web session pending terminal handle recovery', () => { } const keepsPane = { tabs: [expect.objectContaining({ parentTabId: HOST_TAB_ID })] } - it('retires the pane after two authoritative absences', async () => { - await expect(recover(closedTabFrame(3))).resolves.toMatchObject(keepsPane) - await expect(recover(closedTabFrame(4))).resolves.toMatchObject({ tabs: [] }) - }) - - it('retires the pane at once when the frame carries its close record', async () => { - const frame = { - ...closedTabFrame(3), - retiredTerminalSurfaces: [ - { - parentTabId: HOST_TAB_ID, - leafId: LEAF_ID, - terminal: TERMINAL_HANDLE, - ptyId: 'pty-live' - } - ] - } - await expect(recover(frame)).resolves.toMatchObject({ tabs: [] }) - expect(closedPaneCall).not.toHaveBeenCalled() - }) - it('restarts the count when the host publishes the pane again', async () => { await recover(closedTabFrame(3)) await recover({ ...pendingSnapshot(), snapshotVersion: 4 }) await expect(recover(closedTabFrame(5))).resolves.toMatchObject(keepsPane) }) - - it('retires a closed tab whose exited pane an older host still resolves as disconnected', async () => { - closedPaneCall.mockImplementation(async ({ method }) => - method === 'terminal.resolvePane' - ? { - id: 'resolve', - ok: true, - _meta: { runtimeId: 'runtime-1' }, - result: { - terminal: { - handle: TERMINAL_HANDLE, - tabId: HOST_TAB_ID, - leafId: LEAF_ID, - ptyId: 'pty-live', - connected: false, - worktreeId: WORKTREE_ID - } - } - } - : { id: method, ok: false, error: { code: 'unexpected_method', message: method } } - ) - await expect(recover(closedTabFrame(3))).resolves.toMatchObject(keepsPane) - await expect(recover(closedTabFrame(4))).resolves.toMatchObject({ tabs: [] }) - }) - - it('keeps a pane the host still publishes, whatever the pane lookup says', async () => { - for (const snapshotVersion of [3, 4, 5]) { - await expect(recover({ ...pendingSnapshot(), snapshotVersion })).resolves.toMatchObject( - keepsPane - ) - } - }) - - it('never counts a frame the host did not publish', async () => { - await recover({ ...closedTabFrame(3), publicationEpoch: 'none' }) - await expect(recover(closedTabFrame(4))).resolves.toMatchObject(keepsPane) - }) }) }) diff --git a/src/renderer/src/runtime/hosted-terminals/snapshots/remote-runtime-snapshot-outcome.test.ts b/src/renderer/src/runtime/hosted-terminals/snapshots/remote-runtime-snapshot-outcome.test.ts index 4bf0531be7b..7b42e598a03 100644 --- a/src/renderer/src/runtime/hosted-terminals/snapshots/remote-runtime-snapshot-outcome.test.ts +++ b/src/renderer/src/runtime/hosted-terminals/snapshots/remote-runtime-snapshot-outcome.test.ts @@ -172,19 +172,6 @@ describe('remote terminal snapshot outcome reasons', () => { return stream } - it('reports a reply past the client replay limit as permanently unavailable', async () => { - const stream = await subscribeClient() - server.nextRequestedReply = { kind: 'oversized', chunks: 5 } - - await expect(stream.serializeBufferOutcome({ scrollbackRows: 100 })).resolves.toEqual({ - availability: { - kind: 'permanently-unavailable', - reason: 'exceeds-client-replay-limit' - }, - snapshot: null - }) - }) - it('reports a concurrent request as retry-worthy while the legacy path still rejects', async () => { const stream = await subscribeClient() server.nextRequestedReply = { kind: 'hold' } diff --git a/src/renderer/src/runtime/hosts/ssh/runtime-environment-ssh-state.test.ts b/src/renderer/src/runtime/hosts/ssh/runtime-environment-ssh-state.test.ts index 4a9d36ec19a..d90b2cf94a4 100644 --- a/src/renderer/src/runtime/hosts/ssh/runtime-environment-ssh-state.test.ts +++ b/src/renderer/src/runtime/hosts/ssh/runtime-environment-ssh-state.test.ts @@ -72,35 +72,6 @@ beforeEach(() => { }) describe('hydrateRuntimeEnvironmentSshState', () => { - it('prunes removed targets and fetches state only for newly discovered targets', async () => { - const envId = nextEnvId() - useAppStore.getState().setEnvironmentSshTargetsMetadata(envId, [ - { id: 'ssh-keep', label: 'keep' }, - { id: 'ssh-old', label: 'old' } - ]) - useAppStore - .getState() - .setEnvironmentSshConnectionState(envId, 'ssh-keep', connState('ssh-keep')) - useAppStore.getState().setEnvironmentSshConnectionState(envId, 'ssh-old', connState('ssh-old')) - installRpcResponses({ - targets: [ - { id: 'ssh-keep', label: 'keep' }, - { id: 'ssh-new', label: 'new' } - ], - labels: { 'ssh-old': 'old' }, - states: { 'ssh-new': connState('ssh-new', 'connecting') } - }) - - await refreshRuntimeEnvironmentSshTargetMetadata(envId) - - const bucket = useAppStore.getState().sshStateByEnvironment.get(envId) - expect(bucket?.targetLabels.has('ssh-old')).toBe(false) - expect(bucket?.connectionStates.has('ssh-old')).toBe(false) - expect(bucket?.removedTargetLabels.get('ssh-old')).toBe('old') - expect(bucket?.connectionStates.get('ssh-keep')?.status).toBe('connected') - expect(bucket?.connectionStates.get('ssh-new')?.status).toBe('connecting') - }) - it('reruns hydration when a new generation joins an older in-flight request', async () => { const envId = nextEnvId() let resolveOlderTargets!: (value: { targets: { id: string; label: string }[] }) => void @@ -177,27 +148,6 @@ describe('hydrateRuntimeEnvironmentSshState', () => { }) describe('applyRuntimeEnvironmentSshStateChanged', () => { - it('does not touch another environment bucket or local state (no cross-pollution)', () => { - const envA = nextEnvId() - const envB = nextEnvId() - useAppStore.getState().setEnvironmentSshTargetsMetadata(envA, [{ id: 'ssh-1', label: 'a' }]) - useAppStore.getState().setEnvironmentSshTargetsMetadata(envB, [{ id: 'ssh-1', label: 'b' }]) - - applyRuntimeEnvironmentSshStateChanged(envA, 'ssh-1', connState('ssh-1', 'connected')) - - const state = useAppStore.getState() - expect(state.sshStateByEnvironment.get(envA)?.connectionStates.get('ssh-1')?.status).toBe( - 'connected' - ) - const otherBucket = state.sshStateByEnvironment.get(envB) - if (!otherBucket) { - throw new Error('Missing retained SSH environment bucket') - } - expect([...otherBucket.targetLabels]).toEqual([['ssh-1', 'b']]) - expect([...otherBucket.connectionStates]).toEqual([]) - expect([...state.sshConnectionStates]).toEqual([]) - }) - it('re-fetches the authoritative target list for an unknown target instead of trusting the event', async () => { const envId = nextEnvId() // The event races a removal: the authoritative list does not contain it. diff --git a/src/renderer/src/runtime/native-chat/status/structured-agent-session-status-feed.test.ts b/src/renderer/src/runtime/native-chat/status/structured-agent-session-status-feed.test.ts index c71aefd9ac7..8780cdf5fc4 100644 --- a/src/renderer/src/runtime/native-chat/status/structured-agent-session-status-feed.test.ts +++ b/src/renderer/src/runtime/native-chat/status/structured-agent-session-status-feed.test.ts @@ -24,7 +24,6 @@ import { resetStructuredAgentSessionStatusFeedsForTests } from './structured-agent-session-status-feed' -const REMOTE = { kind: 'environment', environmentId: 'env-1' } as const const LOCAL = { kind: 'local' } as const function summary( @@ -51,68 +50,6 @@ function hostEmit(index = 0): (event: AgentSessionStatusEvent) => void { } describe('structured agent session status feed', () => { - it('orders naming observations without changing retained turns, Sleep or confirmation', async () => { - const feed = getStructuredAgentSessionStatusFeed(LOCAL) - const stop = feed.activate() - await vi.advanceTimersByTimeAsync(0) - const retained = { - ...summary('one'), - turnOutcome: 'interruption' as const, - permissionMode: 'plan', - model: 'model', - statusStartedAt: 2, - sleptAt: 3 - } - hostEmit()({ type: 'status', session: retained }) - const status = feed.getSnapshot() - const metadata = (conversationName: string) => ({ - type: 'snapshot' as const, - sessions: [], - namingActiveSessionIds: [], - conversationNames: [{ sessionId: 'one', conversationName }] - }) - hostEmit()(metadata('Before')) - expect(feed.getSnapshot()).toBe(status) - expect(feed.getSessionSleepState('one')).toBe(3) - expect(feed.getSessionSleepState('cold')).toBeUndefined() - expect(feed.getSessionObservation('cold')).toBe('unverifiable') - hostEmit()({ - ...metadata('Before'), - conversationNames: [ - { - sessionId: 'one', - conversationName: 'Before', - namingError: { message: 'Failed', failedAt: 1 } - } - ] - }) - hostEmit()(metadata('Before')) - expect(feed.getNamingSnapshot().get('one')).not.toHaveProperty('namingError') - hostEmit()({ - type: 'snapshot', - sessions: [retained], - conversationNames: JSON.parse('[null, {"sessionId":"one","conversationName":5}]') - }) - expect(feed.getNamingSnapshot().get('one')?.conversationName).toBe('Before') - hostEmit()({ ...metadata('After'), sessions: [{ ...retained, conversationName: 'Stale' }] }) - expect(feed.getNamingSnapshot().get('one')?.conversationName).toBe('After') - hostEmit()({ type: 'end' }) - expect(feed.getSessionObservation('one')).toBe('unverifiable') - expect(feed.getNamingSnapshot().get('one')?.conversationName).toBe('After') - await vi.advanceTimersByTimeAsync(300) - hostEmit(1)({ - ...metadata('Restarted'), - conversationNames: [ - { sessionId: 'one', conversationName: 'Restarted' }, - { sessionId: 'cold', conversationName: 'Cold' } - ] - }) - expect(feed.getSessionObservation('one')).toBe('unverifiable') - expect(feed.getSessionObservation('cold')).toBe('unverifiable') - expect(feed.getSnapshot().get('one')).toMatchObject(retained) - stop() - expect(feed.getNamingSnapshot().get('one')?.conversationName).toBe('Restarted') - }) beforeEach(() => { vi.useFakeTimers() vi.clearAllMocks() @@ -126,44 +63,6 @@ describe('structured agent session status feed', () => { vi.useRealTimers() }) - it('reconnects when the capability probe fails, which is not an answer', async () => { - mocks.supportsCapability.mockRejectedValue(new Error('relay unreachable')) - const feed = getStructuredAgentSessionStatusFeed(REMOTE) - feed.activate() - - await vi.advanceTimersByTimeAsync(0) - expect(mocks.supportsCapability).toHaveBeenCalledTimes(2) - - await vi.advanceTimersByTimeAsync(300) - expect(mocks.supportsCapability).toHaveBeenCalledTimes(4) - expect(mocks.subscribeStatus).not.toHaveBeenCalled() - expect(feed.getSessionSleepState('session-1')).toBeUndefined() - mocks.supportsCapability.mockResolvedValue(true) - await vi.advanceTimersByTimeAsync(500) - hostEmit()({ type: 'snapshot', sessions: [{ ...summary('session-1'), sleptAt: 5 }] }) - expect(feed.getSessionSleepState('session-1')).toBe(5) - }) - - it('keeps Sleep unknown through naming metadata until the full status answer', async () => { - const feed = getStructuredAgentSessionStatusFeed(LOCAL) - feed.activate() - await vi.advanceTimersByTimeAsync(0) - const naming: AgentSessionStatusEvent = { - type: 'snapshot', - sessions: [], - namingActiveSessionIds: [] - } - hostEmit()(naming) - expect(feed.getNamingActive('session-1')).toBe(false) - expect(feed.getSessionSleepState('session-1')).toBeUndefined() - hostEmit()({ type: 'snapshot', sessions: [{ ...summary('session-1'), sleptAt: 5 }] }) - hostEmit()(naming) - expect(feed.getSessionSleepState('session-1')).toBe(5) - hostEmit()({ type: 'snapshot', sessions: [] }) - expect(feed.getSessionSleepState('session-2')).toBeNull() - expect(feed.getSessionSleepState('session-1')).toBe(5) - }) - it('confirms sessions per connection and rejects late events from a disconnected stream', async () => { const feed = getStructuredAgentSessionStatusFeed(LOCAL) const deactivate = feed.activate() @@ -188,31 +87,4 @@ describe('structured agent session status feed', () => { expect(feed.getSessionObservation('one')).toBe('unverifiable') expect(vi.getTimerCount()).toBe(0) }) - - it.each(['error', 'close', 'reject'] as const)( - 'revokes confirmation on subscription %s', - async (failure) => { - let reject!: (error: Error) => void - mocks.subscribeStatus.mockImplementationOnce( - () => - new Promise((_resolve, rejectPromise) => { - reject = rejectPromise - }) - ) - const feed = getStructuredAgentSessionStatusFeed(LOCAL) - feed.activate() - hostEmit()({ type: 'status', session: summary('one') }) - expect(feed.getSessionObservation('one')).toBe('live') - if (failure === 'reject') { - reject(new Error('disconnected')) - } else { - mocks.subscribeStatus.mock.calls[0][failure === 'error' ? 2 : 3]() - } - await vi.advanceTimersByTimeAsync(0) - expect(feed.getSessionObservation('one')).toBe('unverifiable') - hostEmit()({ type: 'status', session: summary('one') }) - expect(feed.getSessionObservation('one')).toBe('unverifiable') - expect(vi.getTimerCount()).toBe(1) - } - ) }) diff --git a/src/renderer/src/runtime/web-session-tabs-sync/identity/web-session-tabs-sync-cross-runtime-worktree.test.ts b/src/renderer/src/runtime/web-session-tabs-sync/identity/web-session-tabs-sync-cross-runtime-worktree.test.ts index 197ed1da624..fa2320d351c 100644 --- a/src/renderer/src/runtime/web-session-tabs-sync/identity/web-session-tabs-sync-cross-runtime-worktree.test.ts +++ b/src/renderer/src/runtime/web-session-tabs-sync/identity/web-session-tabs-sync-cross-runtime-worktree.test.ts @@ -1,8 +1,5 @@ import { beforeEach, describe, expect, it, vi } from 'vitest' -import type { - RuntimeMobileSessionTabsRemovedResult, - RuntimeMobileSessionTabsResult -} from '../../../../../shared/runtime-session-contracts' +import type { RuntimeMobileSessionTabsResult } from '../../../../../shared/runtime-session-contracts' import { applyWebSessionTabsSnapshot } from '../snapshot-api' import type { WebSessionTabsSyncState } from '../state' @@ -157,90 +154,4 @@ describe('a workspace mirrored from two paired runtimes', () => { } } ) - - it("retires the other runtime's own terminals, ready or pending, and refocuses a survivor", () => { - let state = apply(makeState(), makeSnapshot([serverTerminal('agent-1', true)]), SERVER_ENV, NOW) - state = apply( - state, - makeSnapshot([serverTerminal('a-ready', true), serverTerminal('a-pending', false)], { - publicationEpoch: 'desktop-a-epoch' - }), - DESKTOP_A_ENV, - NOW + 1000 - ) - expect(terminalIds(state)).toEqual([ - 'web-terminal-a-pending', - 'web-terminal-a-ready', - 'web-terminal-agent-1' - ]) - state = { ...state, activeTabIdByWorktree: { [WT]: 'web-terminal-a-ready' } } - const retraction: RuntimeMobileSessionTabsRemovedResult = { - ...desktopABrowserSnapshot([], 2), - removed: true, - activeGroupId: null, - activeTabId: null, - activeTabType: null, - tabs: [] - } - state = apply(state, retraction, DESKTOP_A_ENV, NOW + 2000) - expect(terminalIds(state)).toEqual(['web-terminal-agent-1']) - expect(state.activeTabIdByWorktree[WT]).toBe('web-terminal-agent-1') - }) - - it('never keeps a shared terminal bound to a runtime that dropped it', () => { - let state = apply(makeState(), makeSnapshot([serverTerminal('shared', true)]), SERVER_ENV, NOW) - state = apply( - state, - makeSnapshot([serverTerminal('shared', true)], { publicationEpoch: 'desktop-a-epoch' }), - DESKTOP_A_ENV, - NOW + 1000 - ) - expect(state.tabsByWorktree[WT]?.[0]?.ptyId).toBe(`remote:${DESKTOP_A_ENV}@@pty-shared`) - state = apply(state, desktopABrowserSnapshot([], 2), DESKTOP_A_ENV, NOW + 2000) - expect(terminalIds(state)).toEqual([]) - state = apply( - state, - makeSnapshot([serverTerminal('shared', true)], { snapshotVersion: 2 }), - SERVER_ENV, - NOW + 3000 - ) - expect(state.tabsByWorktree[WT]?.[0]?.ptyId).toBe(`remote:${SERVER_ENV}@@pty-shared`) - }) - - it("keeps the server's chats when the other runtime's frame lists none", () => { - const chat = (sessionId: string): RuntimeMobileSessionTabsResult['tabs'][number] => ({ - type: 'agent-session', - id: `agent-session:${sessionId}`, - title: 'Codex Chat', - sessionId, - agent: 'codex', - isActive: false - }) - const chatIds = (s: WebSessionTabsSyncState): string[] => - (s.unifiedTabsByWorktree[WT] ?? []) - .filter((tab) => tab.contentType === 'agent-session') - .map((tab) => tab.entityId) - .sort() - let state = apply( - makeState(), - makeSnapshot([chat('codex-1'), chat('codex-2')]), - SERVER_ENV, - NOW - ) - state = apply(state, desktopABrowserSnapshot([desktopAPage], 1), DESKTOP_A_ENV, NOW + 1000) - expect(chatIds(state)).toEqual(['codex-1', 'codex-2']) - - // A chat both runtimes publish keeps one row, and the server still retires its own. - state = apply( - state, - desktopABrowserSnapshot([desktopAPage, chat('codex-1')], 2), - DESKTOP_A_ENV, - NOW + 2000 - ) - expect((state.unifiedTabsByWorktree[WT] ?? []).map((tab) => tab.id)).toEqual([ - ...new Set((state.unifiedTabsByWorktree[WT] ?? []).map((tab) => tab.id)) - ]) - state = apply(state, makeSnapshot([], { snapshotVersion: 2 }), SERVER_ENV, NOW + 3000) - expect(chatIds(state)).toEqual(['codex-1']) - }) }) diff --git a/src/renderer/src/runtime/web-session-tabs-sync/placement/web-session-open-files-batch-equivalence.test.ts b/src/renderer/src/runtime/web-session-tabs-sync/placement/web-session-open-files-batch-equivalence.test.ts deleted file mode 100644 index 69de3c19874..00000000000 --- a/src/renderer/src/runtime/web-session-tabs-sync/placement/web-session-open-files-batch-equivalence.test.ts +++ /dev/null @@ -1,212 +0,0 @@ -/* Why: the batched open-file path reconciles against a slot-keyed draft instead of - * rebuilding the global array per snapshot. These seeded cases assert it stays - * indistinguishable from repeated single-snapshot reconciliation. */ -import { beforeEach, describe, expect, it, vi } from 'vitest' -import type { RuntimeMobileSessionTabsResult } from '../../../../../shared/runtime-session-contracts' -import type { OpenFile } from '../../../store/slices/editor/types/open-file' -import { resetWebSessionFocusIntentForTests } from '../navigation/web-session-focus-intent' -import { resetWebSessionCloseIntentForTests } from '../closing/web-session-close-intent' -import { resetWebSessionReorderIntentForTests } from '../navigation/web-session-reorder-intent' -import { resetWebAgentSessionHandoffsForTests } from '../../hosted-terminals/creation/agent-launch/web-agent-session-handoff' -import { applyWebSessionTabsSnapshot, applyWebSessionTabsSnapshots } from '../snapshot-api' -import { resetWebSessionTabsSnapshotFreshnessForTests } from '../tracking-lifecycle' -import type { WebSessionTabsSyncState } from '../state' - -vi.mock('../../../store/index', () => ({ useAppStore: { setState: vi.fn() } })) - -const ENV = 'web-env-1' -const NOW = 1_700_000_000_000 -const WORKTREES = ['repo::/w0', 'repo::/w1', 'repo::/w2', 'repo::/bystander'] -const PATHS = ['/repo/a.ts', '/repo/b.ts', '/repo/c.ts', '/repo/d.ts'] -const ENVS: (string | null | undefined)[] = [ENV, 'other-env', null, undefined] - -/** Deterministic LCG so a failing case is reproducible from its seed. */ -function makeRandom(seed: number): () => number { - let value = seed >>> 0 - return () => { - value = (value * 1664525 + 1013904223) >>> 0 - return value / 0x100000000 - } -} - -function resetModuleState(): void { - resetWebSessionTabsSnapshotFreshnessForTests() - resetWebSessionFocusIntentForTests() - resetWebSessionCloseIntentForTests() - resetWebSessionReorderIntentForTests() - resetWebAgentSessionHandoffsForTests() -} - -function baseState(overrides: Partial): WebSessionTabsSyncState { - return { - activeBrowserTabId: null, - activeBrowserTabIdByWorktree: {}, - activeFileId: null, - activeFileIdByWorktree: {}, - activeGroupIdByWorktree: {}, - activeTabId: null, - activeTabIdByWorktree: {}, - activeTabType: 'terminal', - activeTabTypeByWorktree: {}, - activeWorktreeId: WORKTREES[0]!, - agentStatusByPaneKey: {}, - agentStatusEpoch: 0, - browserCertificateFailuresByPageId: {}, - browserPagesByWorkspace: {}, - browserTabsByWorktree: {}, - groupsByWorktree: {}, - layoutByWorktree: {}, - openFiles: [], - editorBuffers: {}, - ptyIdsByTabId: {}, - remoteBrowserPageHandlesByPageId: {}, - tabBarOrderByWorktree: {}, - tabsByWorktree: {}, - terminalLayoutsByTabId: {}, - unifiedTabsByWorktree: {}, - unreadTerminalTabs: {}, - sortEpoch: 0, - ...overrides - } -} - -function pick(random: () => number, items: readonly T[]): T { - return items[Math.floor(random() * items.length)]! -} - -function makeCase(seed: number): { - state: WebSessionTabsSyncState - snapshots: RuntimeMobileSessionTabsResult[] -} { - const random = makeRandom(seed) - const openFiles: OpenFile[] = [] - const fileCount = Math.floor(random() * 9) - for (let i = 0; i < fileCount; i += 1) { - const path = pick(random, PATHS) - const isPreview = random() < 0.2 - openFiles.push({ - id: isPreview ? `markdown-preview::${path}` : path, - filePath: path, - relativePath: path.slice('/repo/'.length), - worktreeId: pick(random, WORKTREES), - language: 'typescript', - isDirty: random() < 0.5, - runtimeEnvironmentId: pick(random, ENVS), - mode: isPreview ? 'markdown-preview' : 'edit', - mirroredFromRuntimeSession: random() < 0.7, - // Why: the mirrored file overwrites every compared field, so only a preserved - // field like this can reveal which duplicate seeded it. - lastKnownDiskSignature: `sig-${seed}-${i}` - } as OpenFile) - } - - const snapshots: RuntimeMobileSessionTabsResult[] = [] - const snapshotCount = 1 + Math.floor(random() * 7) - for (let s = 0; s < snapshotCount; s += 1) { - const worktree = pick(random, WORKTREES) - const tabs: RuntimeMobileSessionTabsResult['tabs'] = [] - const editorCount = Math.floor(random() * 4) - const usedPaths = new Set() - for (let e = 0; e < editorCount; e += 1) { - const path = pick(random, PATHS) - if (usedPaths.has(path)) { - continue - } - usedPaths.add(path) - const relativePath = path.slice('/repo/'.length) - tabs.push( - random() < 0.25 - ? ({ - type: 'markdown', - id: `host-md-${s}-${e}`, - title: relativePath, - filePath: path, - relativePath, - language: 'markdown', - mode: 'markdown-preview', - isDirty: false, - isActive: e === 0, - sourceFileId: path, - sourceFilePath: path, - sourceRelativePath: relativePath, - documentVersion: `file:${path}` - } as RuntimeMobileSessionTabsResult['tabs'][number]) - : ({ - type: 'file', - id: `host-file-${s}-${e}`, - title: relativePath, - filePath: path, - relativePath, - language: 'typescript', - isDirty: random() < 0.5, - isActive: e === 0 - } as RuntimeMobileSessionTabsResult['tabs'][number]) - ) - } - const activeTab = tabs[0] - snapshots.push({ - worktree, - publicationEpoch: 'epoch-1', - snapshotVersion: s + 1, - activeGroupId: `host-group-${worktree}`, - activeTabId: activeTab?.id ?? null, - activeTabType: activeTab ? (activeTab.type === 'markdown' ? 'markdown' : 'file') : null, - tabs - }) - } - - const activeFile = openFiles[Math.floor(random() * Math.max(openFiles.length, 1))] - return { - state: baseState({ - openFiles, - activeFileId: activeFile?.id ?? null, - activeFileIdByWorktree: activeFile ? { [activeFile.worktreeId]: activeFile.id } : {}, - activeTabTypeByWorktree: activeFile ? { [activeFile.worktreeId]: 'editor' } : {}, - activeWorktreeId: pick(random, WORKTREES) - }), - snapshots - } -} - -describe('batched open-file reconciliation', () => { - beforeEach(() => { - resetModuleState() - }) - - it('matches repeated single-snapshot reconciliation for every seeded case', () => { - const divergences: string[] = [] - for (let seed = 1; seed <= 400; seed += 1) { - const { state, snapshots } = makeCase(seed) - const stateCopy = structuredClone(state) - - resetModuleState() - let sequential = state - for (const snapshot of snapshots) { - const patch = applyWebSessionTabsSnapshot(sequential, snapshot, ENV, NOW) - if (patch !== sequential) { - sequential = { ...sequential, ...patch } - } - } - - resetModuleState() - const batched = { - ...state, - ...applyWebSessionTabsSnapshots(state, snapshots, ENV, NOW) - } - - try { - expect(batched.openFiles).toEqual(sequential.openFiles) - expect(batched.activeFileId).toEqual(sequential.activeFileId) - expect(batched.activeFileIdByWorktree).toEqual(sequential.activeFileIdByWorktree) - expect(batched.activeTabTypeByWorktree).toEqual(sequential.activeTabTypeByWorktree) - expect(batched.unifiedTabsByWorktree).toEqual(sequential.unifiedTabsByWorktree) - expect(state).toEqual(stateCopy) - } catch (error) { - divergences.push( - `seed ${seed}: ${error instanceof Error ? error.message.split('\n')[0] : String(error)}` - ) - } - } - expect(divergences).toEqual([]) - }) -}) diff --git a/src/renderer/src/runtime/web-session-tabs-sync/tracking/web-session-tabs-removed-frame-retires-live-publisher.test.ts b/src/renderer/src/runtime/web-session-tabs-sync/tracking/web-session-tabs-removed-frame-retires-live-publisher.test.ts index e5f7493a21e..7e906aa4b9c 100644 --- a/src/renderer/src/runtime/web-session-tabs-sync/tracking/web-session-tabs-removed-frame-retires-live-publisher.test.ts +++ b/src/renderer/src/runtime/web-session-tabs-sync/tracking/web-session-tabs-removed-frame-retires-live-publisher.test.ts @@ -79,34 +79,6 @@ describe('a removal frame must not retire the publisher that is still live', () resetWebSessionTabsSyncTestState() }) - // Bound receipt ledgers by frame, not entry count: one bootstrap inventory shares a frame across worktrees and must not evict its own evidence. - it('keeps every receipt an inventory recorded under one frame, past the bound', () => { - const requestReceivedFrame = nextReceivedSessionTabsFrame() - const worktrees = Array.from( - { length: MAX_TRACKED_SESSION_TABS_RECEIPTS + 64 }, - (_value, index) => `repo::/worktree-${index}` - ) - for (const worktree of worktrees) { - recordReceivedWebSessionTabsSnapshot( - ENVIRONMENT_ID, - { ...liveFrame(1), worktree }, - requestReceivedFrame, - undefined, - 'bootstrap' - ) - } - - for (const worktree of [worktrees[0]!, worktrees.at(-1)!]) { - expect( - shouldApplyRecoveredWebSessionTabsSnapshot( - ENVIRONMENT_ID, - { ...liveFrame(1), worktree }, - requestReceivedFrame - ) - ).toBe(true) - } - }) - // Retraction boundaries must survive ledger bounds or delayed pre-close frames resurrect tabs. it('keeps a worktree fence after enough other worktrees to evict its receipt', () => { const liveReceived = recordReceivedWebSessionTabsSnapshot(ENVIRONMENT_ID, liveFrame(1)) @@ -190,22 +162,4 @@ describe('a removal frame must not retire the publisher that is still live', () shouldApplyRecoveredWebSessionTabsSnapshot(ENVIRONMENT_ID, delayed, delayedReceived) ).toBe(false) }) - - // Exercise full recovery/receipt/decision path for bare and headless-merge epochs. Decision-only tests miss receipt retirement; exact-string differences must not make lockout timing-dependent. - for (const [label, epoch] of [ - ['bare', LIVE_EPOCH], - ['headless-merge', `${LIVE_EPOCH}:headless-merge:abc`] - ] as const) { - it(`readmits a ${label} republication after a removal, through the full path`, () => { - const liveReceived = recordReceivedWebSessionTabsSnapshot(ENVIRONMENT_ID, liveFrame(1)) - expect(admits(liveFrame(1), liveReceived)).toBe(true) - - const removedReceived = recordReceivedWebSessionTabsSnapshot(ENVIRONMENT_ID, removalFrame()) - expect(admits(removalFrame(), removedReceived)).toBe(true) - - const republished = { ...liveFrame(2), publicationEpoch: epoch } - const republishedReceived = recordReceivedWebSessionTabsSnapshot(ENVIRONMENT_ID, republished) - expect(admits(republished, republishedReceived)).toBe(true) - }) - } }) diff --git a/src/renderer/src/runtime/workspace-layout/server/server-held-panes.test.tsx b/src/renderer/src/runtime/workspace-layout/server/server-held-panes.test.tsx index 08c43962271..26791dae64e 100644 --- a/src/renderer/src/runtime/workspace-layout/server/server-held-panes.test.tsx +++ b/src/renderer/src/runtime/workspace-layout/server/server-held-panes.test.tsx @@ -6,7 +6,6 @@ import { createRoot, type Root } from 'react-dom/client' import { afterEach, beforeEach, describe, expect, it, vi, type Mock } from 'vitest' import type * as AgentStatusModule from '@/lib/agent-status/aggregation/worktree-agent-status' import type * as FrameDrawnModule from '../../web-session-tabs-sync/frame-drawn-server-workspaces' -import { makePaneKey } from '../../../../../shared/workspace-layout/stable-pane-id' import { toWebTerminalSurfaceTabId } from '../../../../../shared/workspace-layout/terminal-surface-id' import type { NewerDaemonReopenResult } from '../../../../../shared/newer-daemon-hold' import type { @@ -79,9 +78,7 @@ hoisted.subscribe = streams.subscribe const { setServers, mirroredStore, install, uninstallAll } = createServerHarness(hoisted) const TAB_1 = toWebTerminalSurfaceTabId('host-tab-1') -const PANE_KEY = makePaneKey(TAB_1, LEAF_ID) const REOPENS: Server = { ...CAPABLE, reopensHeldPanes: true } -const FAILED_TITLE = "Couldn't reopen this terminal" /** ENV's workspace with its one pane carrying `status`, as the server publishes it. */ function layoutWith(status: PaneStartStatus | undefined): PublishedWorkspaceLayout { @@ -198,22 +195,4 @@ describe("a layout-owner server's held pane on a paired desktop", () => { /^remote:/ ) }) - - it.each([ - ['refuses', () => Promise.resolve({ reopened: false })], - ['fails', () => Promise.reject(new Error('connection lost'))] - ] as const)('shows the failed state when the server %s', async (_, reply) => { - hoisted.callRuntimeRpc.mockImplementation(reply) - await heldServerPane() - - await act(async () => button('Reopen terminal').click()) - - expect(paneElement.textContent).toContain(FAILED_TITLE) - expect(button('Try again').disabled).toBe(false) - // The banner stays on the failed path while the server still holds the pane. - await publish(layoutWith('held')) - await render() - expect(paneElement.textContent).toContain(FAILED_TITLE) - expect(newerDaemonHoldStore.getState().panes[PANE_KEY]?.phase).toBe('failed') - }) }) diff --git a/src/renderer/src/store/github/requests/provider-request-coalescing.test.ts b/src/renderer/src/store/github/requests/provider-request-coalescing.test.ts index ddb408f741b..2d6482893cf 100644 --- a/src/renderer/src/store/github/requests/provider-request-coalescing.test.ts +++ b/src/renderer/src/store/github/requests/provider-request-coalescing.test.ts @@ -33,57 +33,6 @@ describe('GitHub provider request upgrade coalescing', () => { resetRemoteRuntimeMocks() }) - it('never joins a weaker replacement and never waits out more than one', async () => { - const store = createTestStore() - const weak = Promise.withResolvers>() - const forced = Promise.withResolvers>() - const strict = Promise.withResolvers>() - mockApi.gh.listWorkItems - .mockReturnValueOnce(weak.promise) - .mockReturnValueOnce(forced.promise) - .mockReturnValue(strict.promise) - const first = store.getState().fetchWorkItems('repo-1', '/repo', 24, '') - const forcedFetch = store.getState().fetchWorkItems('repo-1', '/repo', 24, '', { force: true }) - const strictFollowers = Array.from({ length: 2 }, () => - store.getState().fetchWorkItems('repo-1', '/repo', 24, '', { - force: true, - noCache: true, - requireComplete: true - }) - ) - weak.resolve(workItems('weak')) - await first - // Bound upgrade waits so repeating weak replacements cannot starve strict callers, which still share one strict request. - await vi.waitFor(() => expect(mockApi.gh.listWorkItems).toHaveBeenCalledTimes(3)) - expect(mockApi.gh.listWorkItems).toHaveBeenLastCalledWith( - expect.objectContaining({ noCache: true }) - ) - strict.resolve(workItems('strict')) - const strictResults = await Promise.all(strictFollowers) - forced.resolve(workItems('forced')) - await expect(forcedFetch).resolves.toBeUndefined() - - expect(strictResults.map((rows) => rows?.[0].title)).toEqual(['strict', 'strict']) - }) - - it('keeps an invalidated request from removing its replacement dedupe entry', async () => { - const store = createTestStore() - const stale = Promise.withResolvers>() - const fresh = Promise.withResolvers>() - mockApi.gh.listWorkItems.mockReturnValueOnce(stale.promise).mockReturnValue(fresh.promise) - const first = store.getState().fetchWorkItems('repo-1', '/repo', 24, '') - store.getState().evictGitHubRepoCaches('repo-1', '/repo') - const replacement = store.getState().fetchWorkItems('repo-1', '/repo', 24, '', { force: true }) - stale.resolve(workItems('stale')) - await first - const joined = store.getState().fetchWorkItems('repo-1', '/repo', 24, '', { force: true }) - fresh.resolve(workItems('fresh')) - await expect(Promise.all([replacement, joined])).resolves.toEqual([ - workItems('fresh').items, - workItems('fresh').items - ]) - }) - it('rejects partial results for every complete-result waiter and allows a later retry', async () => { const store = createTestStore() const weak = Promise.withResolvers>() diff --git a/src/renderer/src/store/slices/editor/hydration/store-session-editor-hydration.test.ts b/src/renderer/src/store/slices/editor/hydration/store-session-editor-hydration.test.ts deleted file mode 100644 index ef5681a3f34..00000000000 --- a/src/renderer/src/store/slices/editor/hydration/store-session-editor-hydration.test.ts +++ /dev/null @@ -1,272 +0,0 @@ -import { describe, it, expect, vi, beforeEach } from 'vitest' -import type * as AgentStatusModule from '@/lib/agent-status/aggregation/worktree-agent-status' -import { FLOATING_TERMINAL_WORKTREE_ID } from '../../../../../../shared/constants' -import { createTestStore } from '../../../testing/app-store-test-fixture' -import { makeWorktree } from '../../worktrees/worktree.test-fixture' -import { createStoreSessionMockApi } from '../../../testing/app-store-session-api.test-fixture' - -vi.mock('sonner', () => ({ toast: { info: vi.fn(), success: vi.fn(), error: vi.fn() } })) - -vi.mock('@/lib/agent-status/aggregation/worktree-agent-status', async (importOriginal) => { - const actual = await importOriginal() - return { - ...actual, - detectAgentStatusFromTitle: vi.fn().mockReturnValue(null) - } -}) - -createStoreSessionMockApi() - -function ownedEditorFileId( - filePath: string, - worktreeId: string, - runtimeEnvironmentId: string | null | undefined -): string { - const runtimeKey = runtimeEnvironmentId?.trim() || 'local' - return `editor:${encodeURIComponent(worktreeId)}:${encodeURIComponent(runtimeKey)}:${encodeURIComponent(filePath)}` -} - -describe('hydrateEditorSession', () => { - beforeEach(() => { - vi.clearAllMocks() - }) - - it('restores floating workspace markdown files without a repo worktree', () => { - const store = createTestStore() - const filePath = '/orca/userData/floating-workspace/note.md' - const fileId = ownedEditorFileId(filePath, FLOATING_TERMINAL_WORKTREE_ID, null) - - store.setState({ activeWorktreeId: FLOATING_TERMINAL_WORKTREE_ID }) - - store.getState().hydrateEditorSession({ - activeRepoId: null, - activeWorktreeId: FLOATING_TERMINAL_WORKTREE_ID, - activeTabId: null, - tabsByWorktree: {}, - terminalLayoutsByTabId: {}, - openFilesByWorktree: { - [FLOATING_TERMINAL_WORKTREE_ID]: [ - { - filePath, - relativePath: 'note.md', - worktreeId: FLOATING_TERMINAL_WORKTREE_ID, - language: 'markdown', - runtimeEnvironmentId: null, - dirtyDraftContent: '' - } - ] - }, - activeFileIdByWorktree: { - [FLOATING_TERMINAL_WORKTREE_ID]: '/orca/userData/floating-workspace/note.md' - }, - activeTabTypeByWorktree: { [FLOATING_TERMINAL_WORKTREE_ID]: 'editor' } - }) - - const s = store.getState() - expect(s.openFiles).toEqual([ - expect.objectContaining({ - id: fileId, - filePath, - worktreeId: FLOATING_TERMINAL_WORKTREE_ID, - runtimeEnvironmentId: null, - isDirty: true - }) - ]) - expect(s.editorDrafts).toEqual({ [fileId]: '' }) - expect(s.markdownFrontmatterVisible).toEqual({}) - expect(s.activeFileIdByWorktree[FLOATING_TERMINAL_WORKTREE_ID]).toBe(fileId) - }) - - it('keeps same-path local and runtime legacy references on their original owners', () => { - const store = createTestStore() - const wt = 'repo1::/path/wt1' - const filePath = '/path/wt1/src/app.ts' - const runtimeEnvironmentId = 'runtime-1' - const runtimeFileId = ownedEditorFileId(filePath, wt, runtimeEnvironmentId) - const groupId = 'group-same-path-owners' - - store.setState({ - repos: [ - { id: 'repo1', path: '/repo1', displayName: 'Repo 1', badgeColor: '#000', addedAt: 0 } - ], - worktreesByRepo: { - repo1: [makeWorktree({ id: wt, repoId: 'repo1', path: '/path/wt1' })] - }, - activeWorktreeId: wt - }) - - const session = { - activeRepoId: 'repo1', - activeWorktreeId: wt, - activeTabId: null, - tabsByWorktree: {}, - terminalLayoutsByTabId: {}, - openFilesByWorktree: { - [wt]: [ - { - filePath, - relativePath: 'src/app.ts', - worktreeId: wt, - language: 'typescript' - }, - { - filePath, - relativePath: 'src/app.ts', - worktreeId: wt, - language: 'typescript', - runtimeEnvironmentId - } - ] - }, - activeFileIdByWorktree: { [wt]: filePath }, - activeTabTypeByWorktree: { [wt]: 'editor' as const }, - unifiedTabs: { - [wt]: [ - { - id: filePath, - entityId: filePath, - groupId, - worktreeId: wt, - contentType: 'editor' as const, - label: 'app.ts', - customLabel: null, - color: null, - sortOrder: 0, - createdAt: 1 - }, - { - id: runtimeFileId, - entityId: runtimeFileId, - groupId, - worktreeId: wt, - contentType: 'editor' as const, - label: 'app.ts', - customLabel: null, - color: null, - sortOrder: 1, - createdAt: 2 - } - ] - }, - tabGroups: { - [wt]: [ - { - id: groupId, - worktreeId: wt, - activeTabId: filePath, - tabOrder: [filePath, runtimeFileId], - recentTabIds: [runtimeFileId, filePath] - } - ] - }, - activeGroupIdByWorktree: { [wt]: groupId } - } - - store.getState().hydrateTabsSession(session) - store.getState().hydrateEditorSession(session) - - const s = store.getState() - expect(s.openFiles).toEqual([ - expect.objectContaining({ - id: filePath, - filePath, - worktreeId: wt, - runtimeEnvironmentId: undefined - }), - expect.objectContaining({ - id: runtimeFileId, - filePath, - worktreeId: wt, - runtimeEnvironmentId - }) - ]) - expect(s.activeFileIdByWorktree[wt]).toBe(filePath) - expect(s.unifiedTabsByWorktree[wt]?.map((tab) => tab.id)).toEqual([filePath, runtimeFileId]) - expect(s.unifiedTabsByWorktree[wt]?.map((tab) => tab.entityId)).toEqual([ - filePath, - runtimeFileId - ]) - expect(s.groupsByWorktree[wt]?.[0]).toEqual( - expect.objectContaining({ - activeTabId: filePath, - tabOrder: [filePath, runtimeFileId], - recentTabIds: [runtimeFileId, filePath] - }) - ) - }) - - it('migrates legacy floating unified tab file-path references to the hydrated owner id', () => { - const store = createTestStore() - const filePath = '/orca/userData/floating-workspace/README.md' - const fileId = ownedEditorFileId(filePath, FLOATING_TERMINAL_WORKTREE_ID, null) - const groupId = 'floating-group-legacy' - - store.setState({ activeWorktreeId: FLOATING_TERMINAL_WORKTREE_ID }) - - const session = { - activeRepoId: null, - activeWorktreeId: FLOATING_TERMINAL_WORKTREE_ID, - activeTabId: null, - tabsByWorktree: {}, - terminalLayoutsByTabId: {}, - openFilesByWorktree: { - [FLOATING_TERMINAL_WORKTREE_ID]: [ - { - filePath, - relativePath: 'README.md', - worktreeId: FLOATING_TERMINAL_WORKTREE_ID, - language: 'markdown', - runtimeEnvironmentId: null - } - ] - }, - activeFileIdByWorktree: { [FLOATING_TERMINAL_WORKTREE_ID]: filePath }, - activeTabTypeByWorktree: { [FLOATING_TERMINAL_WORKTREE_ID]: 'editor' as const }, - unifiedTabs: { - [FLOATING_TERMINAL_WORKTREE_ID]: [ - { - id: filePath, - entityId: filePath, - groupId, - worktreeId: FLOATING_TERMINAL_WORKTREE_ID, - contentType: 'editor' as const, - label: 'README.md', - customLabel: null, - color: null, - sortOrder: 0, - createdAt: 1 - } - ] - }, - tabGroups: { - [FLOATING_TERMINAL_WORKTREE_ID]: [ - { - id: groupId, - worktreeId: FLOATING_TERMINAL_WORKTREE_ID, - activeTabId: filePath, - tabOrder: [filePath], - recentTabIds: [filePath] - } - ] - }, - activeGroupIdByWorktree: { [FLOATING_TERMINAL_WORKTREE_ID]: groupId } - } - - store.getState().hydrateTabsSession(session) - store.getState().hydrateEditorSession(session) - - const s = store.getState() - expect(s.openFiles[0]?.id).toBe(fileId) - expect(s.activeFileIdByWorktree[FLOATING_TERMINAL_WORKTREE_ID]).toBe(fileId) - expect(s.unifiedTabsByWorktree[FLOATING_TERMINAL_WORKTREE_ID]?.[0]).toEqual( - expect.objectContaining({ id: fileId, entityId: fileId }) - ) - expect(s.groupsByWorktree[FLOATING_TERMINAL_WORKTREE_ID]?.[0]).toEqual( - expect.objectContaining({ - activeTabId: fileId, - tabOrder: [fileId], - recentTabIds: [fileId] - }) - ) - }) -}) diff --git a/src/renderer/src/store/slices/ssh/ssh.test.ts b/src/renderer/src/store/slices/ssh/ssh.test.ts deleted file mode 100644 index 5006d4416fc..00000000000 --- a/src/renderer/src/store/slices/ssh/ssh.test.ts +++ /dev/null @@ -1,214 +0,0 @@ -import { describe, expect, it } from 'vitest' -import { toAppSshPtyId } from '../../../../../shared/ssh-pty-id' -import { createTestStore } from '../../testing/app-store-test-fixture' -import { makeTab } from '../../terminals/terminal-tab.test-fixture' -import { makeWorktree } from '../worktrees/worktree.test-fixture' -import { TEST_REPO } from '../../repos/repo.test-fixture' - -describe('createSshSlice', () => { - it('clears renderer state and deferred reconnect metadata for a removed SSH target', () => { - const store = createTestStore() - const targetId = 'ssh-1' - const otherTargetId = 'ssh-2' - const worktreeId = 'repo-ssh::/remote/project' - const otherWorktreeId = 'repo-other::/remote/project' - const removedPtyId = toAppSshPtyId(targetId, 'pty-live') - const staleLastKnownPtyId = toAppSshPtyId(targetId, 'pty-last-known') - const otherPtyId = toAppSshPtyId(otherTargetId, 'pty-other') - - store.setState({ - repos: [ - { ...TEST_REPO, id: 'repo-ssh', connectionId: targetId }, - { ...TEST_REPO, id: 'repo-other', connectionId: otherTargetId } - ], - worktreesByRepo: { - 'repo-ssh': [makeWorktree({ id: worktreeId, repoId: 'repo-ssh' })], - 'repo-other': [makeWorktree({ id: otherWorktreeId, repoId: 'repo-other' })] - }, - tabsByWorktree: { - [worktreeId]: [ - makeTab({ - id: 'tab-ssh', - worktreeId, - ptyId: removedPtyId, - pendingActivationSpawn: true - }) - ], - [otherWorktreeId]: [ - makeTab({ id: 'tab-other', worktreeId: otherWorktreeId, ptyId: otherPtyId }), - makeTab({ id: 'tab-stale-last-known', worktreeId: otherWorktreeId }) - ] - }, - ptyIdsByTabId: { - 'tab-ssh': [removedPtyId], - 'tab-other': [otherPtyId] - }, - lastKnownRelayPtyIdByTabId: { - 'tab-ssh': removedPtyId, - 'tab-stale-last-known': staleLastKnownPtyId, - 'tab-other': otherPtyId - }, - pendingPaneAccountRestarts: { - [removedPtyId]: 'codex', - [otherPtyId]: 'codex' - }, - codexRestartNoticeByPtyId: { - [removedPtyId]: { previousAccountLabel: 'Old', nextAccountLabel: 'New' }, - [otherPtyId]: { previousAccountLabel: 'Other old', nextAccountLabel: 'Other new' } - }, - sshConnectionStates: new Map([ - [targetId, { targetId, status: 'disconnected', error: null, reconnectAttempt: 0 }], - [ - otherTargetId, - { targetId: otherTargetId, status: 'connected', error: null, reconnectAttempt: 0 } - ] - ]), - sshTargetLabels: new Map([ - [targetId, 'Removed target'], - [otherTargetId, 'Other target'] - ]), - portForwardsByConnection: { - [targetId]: [ - { - id: 'pf-1', - connectionId: targetId, - localPort: 3000, - remoteHost: '127.0.0.1', - remotePort: 3000 - } - ], - [otherTargetId]: [ - { - id: 'pf-2', - connectionId: otherTargetId, - localPort: 4000, - remoteHost: '127.0.0.1', - remotePort: 4000 - } - ] - }, - detectedPortsByConnection: { - [targetId]: [{ port: 3000, host: '127.0.0.1' }], - [otherTargetId]: [{ port: 4000, host: '127.0.0.1' }] - }, - sshCredentialQueue: [ - { requestId: 'req-1', targetId, kind: 'passphrase', detail: 'key' }, - { requestId: 'req-2', targetId: otherTargetId, kind: 'password', detail: 'password' } - ], - deferredSshReconnectTargets: [targetId, otherTargetId], - transientClearedAgentStatusConnectionIds: { - [targetId]: true, - [otherTargetId]: true - }, - deferredSshSessionIdsByTabId: { - 'tab-ssh': 'legacy-session-without-target-prefix', - 'tab-stale-encoded': toAppSshPtyId(targetId, 'pty-1'), - 'tab-other': toAppSshPtyId(otherTargetId, 'pty-2') - }, - // Pending reconnect counts as orphan-sweep liveness; removal must clear it or pin the dead tab (#9911). - pendingReconnectPtyIdByTabId: { - 'tab-ssh': toAppSshPtyId(targetId, 'pty-1'), - 'tab-stale-encoded': toAppSshPtyId(targetId, 'pty-9'), - 'tab-other': toAppSshPtyId(otherTargetId, 'pty-2') - } - }) - - store.getState().clearRemovedSshTargetState(targetId) - - const state = store.getState() - expect(state.sshConnectionStates.has(targetId)).toBe(false) - expect(state.sshTargetLabels.has(targetId)).toBe(false) - expect(state.portForwardsByConnection[targetId]).toBeUndefined() - expect(state.detectedPortsByConnection[targetId]).toBeUndefined() - expect(state.sshCredentialQueue.map((req) => req.targetId)).toEqual([otherTargetId]) - expect(state.deferredSshReconnectTargets).toEqual([otherTargetId]) - expect(state.transientClearedAgentStatusConnectionIds).toEqual({ - [otherTargetId]: true - }) - expect(state.deferredSshSessionIdsByTabId).toEqual({ - 'tab-other': toAppSshPtyId(otherTargetId, 'pty-2') - }) - expect(state.pendingReconnectPtyIdByTabId).toEqual({ - 'tab-other': toAppSshPtyId(otherTargetId, 'pty-2') - }) - expect(state.tabsByWorktree[worktreeId][0]).toMatchObject({ id: 'tab-ssh', ptyId: null }) - expect('pendingActivationSpawn' in state.tabsByWorktree[worktreeId][0]).toBe(false) - expect(state.ptyIdsByTabId['tab-ssh']).toEqual([]) - expect(state.lastKnownRelayPtyIdByTabId['tab-ssh']).toBeUndefined() - expect(state.lastKnownRelayPtyIdByTabId['tab-stale-last-known']).toBeUndefined() - expect(state.pendingPaneAccountRestarts[removedPtyId]).toBeUndefined() - expect(state.codexRestartNoticeByPtyId[removedPtyId]).toBeUndefined() - - expect(state.sshConnectionStates.get(otherTargetId)?.status).toBe('connected') - expect(state.sshTargetLabels.get(otherTargetId)).toBe('Other target') - expect(state.portForwardsByConnection[otherTargetId]).toHaveLength(1) - expect(state.detectedPortsByConnection[otherTargetId]).toHaveLength(1) - expect(state.tabsByWorktree[otherWorktreeId][0]?.ptyId).toBe(otherPtyId) - expect(state.ptyIdsByTabId['tab-other']).toEqual([otherPtyId]) - expect(state.lastKnownRelayPtyIdByTabId['tab-other']).toBe(otherPtyId) - expect(state.pendingPaneAccountRestarts[otherPtyId]).toBe('codex') - expect(state.codexRestartNoticeByPtyId[otherPtyId]).toEqual({ - previousAccountLabel: 'Other old', - nextAccountLabel: 'Other new' - }) - }) - - it("preserves another SSH target's terminal bindings when repo ids collide", () => { - const store = createTestStore() - const removedTargetId = 'target-a' - const survivingTargetId = 'target-b' - const removedWorktreeId = 'shared::/srv/a/work' - const survivingWorktreeId = 'shared::/srv/b/work' - const removedPtyId = toAppSshPtyId(removedTargetId, 'pty-a') - const survivingPtyId = toAppSshPtyId(survivingTargetId, 'pty-b') - - store.setState({ - repos: [ - { - ...TEST_REPO, - id: 'shared', - connectionId: removedTargetId, - executionHostId: 'ssh:target-a' - }, - { - ...TEST_REPO, - id: 'shared', - connectionId: survivingTargetId, - executionHostId: 'ssh:target-b' - } - ], - worktreesByRepo: { - shared: [ - makeWorktree({ - id: removedWorktreeId, - repoId: 'shared', - hostId: 'ssh:target-a' - }), - makeWorktree({ - id: survivingWorktreeId, - repoId: 'shared', - hostId: 'ssh:target-b' - }) - ] - }, - tabsByWorktree: { - [removedWorktreeId]: [ - makeTab({ id: 'tab-a', worktreeId: removedWorktreeId, ptyId: removedPtyId }) - ], - [survivingWorktreeId]: [ - makeTab({ id: 'tab-b', worktreeId: survivingWorktreeId, ptyId: survivingPtyId }) - ] - }, - ptyIdsByTabId: { 'tab-a': [removedPtyId], 'tab-b': [survivingPtyId] }, - lastKnownRelayPtyIdByTabId: { 'tab-a': removedPtyId, 'tab-b': survivingPtyId } - }) - - store.getState().clearRemovedSshTargetState(removedTargetId) - - const state = store.getState() - expect(state.tabsByWorktree[removedWorktreeId][0]?.ptyId).toBeNull() - expect(state.tabsByWorktree[survivingWorktreeId][0]?.ptyId).toBe(survivingPtyId) - expect(state.ptyIdsByTabId['tab-b']).toEqual([survivingPtyId]) - expect(state.lastKnownRelayPtyIdByTabId['tab-b']).toBe(survivingPtyId) - }) -}) diff --git a/src/renderer/src/store/slices/worktrees/session/worktrees-terminal-pr-url-linking.test.ts b/src/renderer/src/store/slices/worktrees/session/worktrees-terminal-pr-url-linking.test.ts index 07de1ff300f..050883c6bf5 100644 --- a/src/renderer/src/store/slices/worktrees/session/worktrees-terminal-pr-url-linking.test.ts +++ b/src/renderer/src/store/slices/worktrees/session/worktrees-terminal-pr-url-linking.test.ts @@ -61,51 +61,6 @@ describe('worktree remote runtime mutations', () => { }) }) - it('rechecks GitHub PR suppression after branch confirmation resolves', async () => { - const store = createTestStore() - let resolveLookup: (value: HostedReviewInfo | null) => void = () => {} - installReviewRead().mockImplementation( - () => - new Promise((resolve) => { - resolveLookup = resolve - }) - ) - const wt = makeWorktree({ - id: 'repo1::/path/wt1', - repoId: 'repo1', - path: '/worktrees/orca', - branch: 'refs/heads/feature/pr-link', - linkedPR: null, - pushTarget: { - remoteName: 'origin', - branchName: 'feature/pr-link' - } - }) - store.setState({ - repos: [ - { id: 'repo1', path: '/repos/orca', displayName: 'orca', badgeColor: '#000', addedAt: 0 } - ], - worktreesByRepo: { repo1: [wt] } - }) - - store.getState().observeTerminalGitHubPullRequestLink(wt.id, { - url: 'https://github.com/acme/orca/pull/42', - slug: { owner: 'acme', repo: 'orca' }, - number: 42 - }) - await vi.dynamicImportSettled() - store.setState({ - worktreesByRepo: { - repo1: [{ ...wt, linkedPR: null, suppressedGitHubPR: 42 }] - } - }) - - resolveLookup(confirmedReview(42)) - await vi.dynamicImportSettled() - - expect(mockApi.worktrees.updateMeta).not.toHaveBeenCalled() - }) - it('preserves a manual primary while appending a PR confirmed after the selection changes', async () => { const store = createTestStore() let resolveLookup: (value: HostedReviewInfo | null) => void = () => {} @@ -153,108 +108,4 @@ describe('worktree remote runtime mutations', () => { ).toEqual([7, 42]) expect(mockApi.worktrees.updateMeta).toHaveBeenCalled() }) - - it('appends a confirmed PR without overwriting a primary changed during push target lookup', async () => { - const store = createTestStore() - installReviewRead().mockResolvedValue(confirmedReview(42)) - let resolvePushTarget: (value: { - baseBranch: string - pushTarget: { remoteName: string; branchName: string } - }) => void = () => {} - mockApi.worktrees.resolvePrBase.mockImplementationOnce( - () => - new Promise<{ - baseBranch: string - pushTarget: { remoteName: string; branchName: string } - }>((resolve) => { - resolvePushTarget = resolve - }) - ) - const wt = makeWorktree({ - id: 'repo1::/path/wt1', - repoId: 'repo1', - path: '/worktrees/orca', - branch: 'refs/heads/feature/pr-link' - }) - store.setState({ - repos: [ - { id: 'repo1', path: '/repos/orca', displayName: 'orca', badgeColor: '#000', addedAt: 0 } - ], - worktreesByRepo: { repo1: [wt] } - }) - - store.getState().observeTerminalGitHubPullRequestLink(wt.id, { - url: 'https://github.com/acme/orca/pull/42', - slug: { owner: 'acme', repo: 'orca' }, - number: 42 - }) - - await vi.dynamicImportSettled() - - expect(mockApi.worktrees.resolvePrBase).toHaveBeenCalledWith({ - repoId: 'repo1', - prNumber: 42 - }) - - store.setState({ - worktreesByRepo: { repo1: [{ ...wt, linkedPR: 7 }] } - }) - - resolvePushTarget({ - baseBranch: 'main', - pushTarget: { remoteName: 'origin', branchName: 'feature/pr-link' } - }) - await vi.dynamicImportSettled() - - expect(store.getState().worktreesByRepo.repo1[0]?.linkedPR).toBe(7) - expect( - getWorkspaceAttachments(store.getState().worktreesByRepo.repo1[0]).map((item) => item.number) - ).toEqual([7, 42]) - expect(mockApi.worktrees.updateMeta).toHaveBeenCalled() - }) - - it('rejects a same-number branch result for a different repository URL', async () => { - const store = createTestStore() - const readHostedReview = installReviewRead().mockResolvedValue(confirmedReview(42)) - const wt = makeWorktree({ - id: 'repo1::/path/wt1', - repoId: 'repo1', - path: '/worktrees/orca', - branch: 'refs/heads/feature/pr-link' - }) - mockApi.worktrees.resolvePrBase.mockResolvedValueOnce({ baseBranch: 'main' }) - store.setState({ - repos: [ - { id: 'repo1', path: '/repos/orca', displayName: 'orca', badgeColor: '#000', addedAt: 0 } - ], - worktreesByRepo: { repo1: [wt] } - }) - - store.getState().observeTerminalGitHubPullRequestLink(wt.id, { - url: 'https://github.com/acme/docs/pull/42', - slug: { owner: 'acme', repo: 'docs' }, - number: 42 - }) - - expect(store.getState().worktreesByRepo.repo1[0]?.linkedPR).toBeNull() - await vi.waitFor(() => expect(readHostedReview).toHaveBeenCalled()) - expect(readHostedReview).toHaveBeenCalledWith({ - repoPath: '/repos/orca', - branch: 'feature/pr-link', - force: true, - repoId: 'repo1', - active: true, - repoOwnerExecutionHostId: 'local', - currentHeadOid: wt.head, - linkedGitHubPR: null, - linkedGitLabMR: null, - linkedBitbucketPR: null, - linkedAzureDevOpsPR: null, - linkedGiteaPR: null - }) - - await vi.dynamicImportSettled() - - expect(mockApi.worktrees.updateMeta).not.toHaveBeenCalled() - }) }) diff --git a/src/renderer/src/store/slices/worktrees/teardown/worktrees-purge-terminal-state.test.ts b/src/renderer/src/store/slices/worktrees/teardown/worktrees-purge-terminal-state.test.ts deleted file mode 100644 index 762e008f1c6..00000000000 --- a/src/renderer/src/store/slices/worktrees/teardown/worktrees-purge-terminal-state.test.ts +++ /dev/null @@ -1,241 +0,0 @@ -import { beforeEach, describe, expect, it, vi } from 'vitest' -import type { AppState } from '../../../types' -import { FLOATING_TERMINAL_WORKTREE_ID } from '../../../../../../shared/constants' -import { makeLineage } from '../metadata/workspace-lineage.test-fixture' -import { - createTestStore, - resetRemoteRuntimeMocks, - resetWorktreeSliceModuleMemory -} from '../worktrees-slice-test-harness' - -const requestWorktreeBaseFallbackNotice = vi.hoisted(() => vi.fn()) - -vi.mock('sonner', () => ({ - toast: { - warning: vi.fn(), - info: vi.fn(), - success: vi.fn(), - error: vi.fn(), - dismiss: vi.fn() - } -})) - -vi.mock('@/components/worktree-base-fallback-notice', () => ({ - requestWorktreeBaseFallbackNotice -})) - -beforeEach(resetWorktreeSliceModuleMemory) - -// Why: design §4.4 — purgeWorktreeTerminalState wipes every worktree-scoped map symmetrically so no entry is stranded. -describe('purgeWorktreeTerminalState direct (design §4.4)', () => { - beforeEach(() => { - vi.clearAllMocks() - resetRemoteRuntimeMocks() - }) - - it('wipes tab-id-keyed maps (terminalLayoutsByTabId, ptyIdsByTabId) and clears actives', () => { - const store = createTestStore() - - store.setState({ - tabsByWorktree: { - 'repoA::/a/wt1': [ - { id: 'tab-1', worktreeId: 'repoA::/a/wt1' }, - { id: 'tab-2', worktreeId: 'repoA::/a/wt1' } - ], - 'repoA::/a/wt2': [{ id: 'tab-3', worktreeId: 'repoA::/a/wt2' }] - }, - terminalLayoutsByTabId: { - 'tab-1': { panes: [] }, - 'tab-2': { panes: [] }, - 'tab-3': { panes: [] } - }, - ptyIdsByTabId: { 'tab-1': ['pty-1'], 'tab-2': ['pty-2'], 'tab-3': ['pty-3'] }, - expandedPaneByTabId: { 'tab-1': true, 'tab-2': false, 'tab-3': true }, - canExpandPaneByTabId: { 'tab-1': true, 'tab-2': true, 'tab-3': false }, - runtimePaneTitlesByTabId: { 'tab-1': 'claude', 'tab-3': 'bash' }, - automaticAgentResumeClaimsByTabId: { - 'tab-1': { - worktreeId: 'repoA::/a/wt1', - launchAgent: 'codex', - providerSession: { key: 'session_id', id: 'sess-1' } - }, - 'tab-3': { - worktreeId: 'repoA::/a/wt2', - launchAgent: 'codex', - providerSession: { key: 'session_id', id: 'sess-3' } - } - }, - openFiles: [ - { - id: 'file-1', - worktreeId: 'repoA::/a/wt1', - filePath: '/a/wt1/a.ts', - relativePath: 'a.ts', - language: 'typescript', - isDirty: false, - markdownPreviewSourceFileId: 'source-1', - isPreview: false, - mode: 'edit' as const - } - ], - editorDrafts: { 'file-1': 'draft', 'file-99': 'other' }, - markdownFrontmatterVisible: { 'source-1': true, 'file-99': true }, - gitIgnoredPathsByWorktree: { - 'repoA::/a/wt1': ['dist/'], - 'repoA::/a/wt2': ['coverage/'] - }, - gitStatusHeadByWorktree: { - 'repoA::/a/wt1': 'head-1', - 'repoA::/a/wt2': 'head-2' - }, - gitBranchLineTotalByWorktree: { - 'repoA::/a/wt1': { added: 24, removed: 3, mergeBase: 'base-1' }, - 'repoA::/a/wt2': { added: 1, removed: 0, mergeBase: 'base-2' } - }, - gitBranchCompareRequestStatusHeadByWorktree: { - 'repoA::/a/wt1': 'head-1', - 'repoA::/a/wt2': 'head-2' - }, - rightSidebarTabByWorktree: { - 'repoA::/a/wt1': 'search' as never, - 'repoA::/a/wt2': 'checks' - }, - activeWorktreeId: 'repoA::/a/wt1', - worktreeLineageById: { - 'repoA::/a/wt1': makeLineage({ worktreeId: 'repoA::/a/wt1' }), - 'repoA::/a/wt2': makeLineage({ worktreeId: 'repoA::/a/wt2' }) - }, - activeFileId: 'file-1', - activeTabId: 'tab-1', - activeTabType: 'editor' as const - } as unknown as Partial) - - store.getState().purgeWorktreeTerminalState(['repoA::/a/wt1']) - - const s = store.getState() - expect(s.tabsByWorktree).toEqual({ - 'repoA::/a/wt2': [{ id: 'tab-3', worktreeId: 'repoA::/a/wt2' }] - }) - expect(s.worktreeLineageById).toEqual({ - 'repoA::/a/wt2': makeLineage({ worktreeId: 'repoA::/a/wt2' }) - }) - expect(s.terminalLayoutsByTabId).toEqual({ 'tab-3': { panes: [] } }) - expect(s.ptyIdsByTabId).toEqual({ 'tab-3': ['pty-3'] }) - expect(s.expandedPaneByTabId).toEqual({ 'tab-3': true }) - expect(s.canExpandPaneByTabId).toEqual({ 'tab-3': false }) - expect(s.runtimePaneTitlesByTabId).toEqual({ 'tab-3': 'bash' }) - expect(s.automaticAgentResumeClaimsByTabId).toEqual({ - 'tab-3': { - worktreeId: 'repoA::/a/wt2', - launchAgent: 'codex', - providerSession: { key: 'session_id', id: 'sess-3' } - } - }) - expect(s.openFiles).toEqual([]) - expect(s.editorDrafts).toEqual({ 'file-99': 'other' }) - expect(s.markdownFrontmatterVisible).toEqual({ 'file-99': true }) - expect(s.gitStatusHeadByWorktree).toEqual({ 'repoA::/a/wt2': 'head-2' }) - expect(s.gitBranchLineTotalByWorktree).toEqual({ - 'repoA::/a/wt2': { added: 1, removed: 0, mergeBase: 'base-2' } - }) - expect(s.gitBranchCompareRequestStatusHeadByWorktree).toEqual({ - 'repoA::/a/wt2': 'head-2' - }) - expect(s.gitIgnoredPathsByWorktree).toEqual({ 'repoA::/a/wt2': ['coverage/'] }) - expect(s.rightSidebarTabByWorktree).toEqual({ 'repoA::/a/wt2': 'checks' }) - expect(s.activeWorktreeId).toBeNull() - expect(s.activeFileId).toBeNull() - expect(s.activeTabId).toBeNull() - expect(s.activeTabType).toBe('terminal') - }) - - it('ignores the floating workspace sentinel while purging mixed real ids', () => { - const store = createTestStore() - const staleId = 'repoA::/a/wt1' - const floatingFile = { - id: 'floating-file', - worktreeId: FLOATING_TERMINAL_WORKTREE_ID, - filePath: '/floating/note.md', - relativePath: 'note.md', - language: 'markdown', - isDirty: false, - isPreview: false, - mode: 'edit' as const - } - - store.setState({ - tabsByWorktree: { - [staleId]: [{ id: 'tab-1', worktreeId: staleId }], - [FLOATING_TERMINAL_WORKTREE_ID]: [ - { id: 'floating-terminal-tab', worktreeId: FLOATING_TERMINAL_WORKTREE_ID } - ] - }, - browserTabsByWorktree: { - [FLOATING_TERMINAL_WORKTREE_ID]: [{ id: 'floating-browser', url: 'https://orca.test' }] - }, - openFiles: [ - floatingFile, - { - id: 'stale-file', - worktreeId: staleId, - filePath: '/a/wt1/stale.ts', - relativePath: 'stale.ts', - language: 'typescript', - isDirty: false, - isPreview: false, - mode: 'edit' as const - } - ], - activeFileIdByWorktree: { - [FLOATING_TERMINAL_WORKTREE_ID]: 'floating-file', - [staleId]: 'stale-file' - }, - unifiedTabsByWorktree: { - [FLOATING_TERMINAL_WORKTREE_ID]: [ - { - id: 'floating-unified-tab', - type: 'editor', - worktreeId: FLOATING_TERMINAL_WORKTREE_ID, - fileId: 'floating-file' - } - ], - [staleId]: [{ id: 'stale-unified-tab', worktreeId: staleId }] - }, - groupsByWorktree: { - [FLOATING_TERMINAL_WORKTREE_ID]: [ - { - id: 'floating-group', - worktreeId: FLOATING_TERMINAL_WORKTREE_ID, - activeTabId: 'floating-unified-tab' - } - ], - [staleId]: [{ id: 'stale-group', worktreeId: staleId, activeTabId: 'stale-unified-tab' }] - }, - layoutByWorktree: { - [FLOATING_TERMINAL_WORKTREE_ID]: { type: 'leaf', groupId: 'floating-group' }, - [staleId]: { type: 'leaf', groupId: 'stale-group' } - } - } as unknown as Partial) - - store.getState().purgeWorktreeTerminalState([staleId, FLOATING_TERMINAL_WORKTREE_ID]) - - expect(store.getState().tabsByWorktree).toEqual({ - [FLOATING_TERMINAL_WORKTREE_ID]: [ - { id: 'floating-terminal-tab', worktreeId: FLOATING_TERMINAL_WORKTREE_ID } - ] - }) - expect(store.getState().browserTabsByWorktree[FLOATING_TERMINAL_WORKTREE_ID]).toEqual([ - { id: 'floating-browser', url: 'https://orca.test' } - ]) - expect(store.getState().openFiles).toEqual([floatingFile]) - expect(store.getState().activeFileIdByWorktree).toEqual({ - [FLOATING_TERMINAL_WORKTREE_ID]: 'floating-file' - }) - expect(store.getState().unifiedTabsByWorktree[FLOATING_TERMINAL_WORKTREE_ID]).toHaveLength(1) - expect(store.getState().groupsByWorktree[FLOATING_TERMINAL_WORKTREE_ID]).toHaveLength(1) - expect(store.getState().layoutByWorktree[FLOATING_TERMINAL_WORKTREE_ID]).toEqual({ - type: 'leaf', - groupId: 'floating-group' - }) - }) -}) diff --git a/src/renderer/src/store/terminals/store-session-workspace-hydration.test.ts b/src/renderer/src/store/terminals/store-session-workspace-hydration.test.ts deleted file mode 100644 index d8e1856b6db..00000000000 --- a/src/renderer/src/store/terminals/store-session-workspace-hydration.test.ts +++ /dev/null @@ -1,250 +0,0 @@ -import { describe, it, expect, vi, beforeEach } from 'vitest' -import type * as AgentStatusModule from '@/lib/agent-status/aggregation/worktree-agent-status' -import type { DetectedWorktreeListResult, Worktree } from '../../../../shared/worktree/types' -import { getDefaultWorkspaceSession } from '../../../../shared/constants' -import { folderWorkspaceKey } from '../../../../shared/workspace-scope' -import { createTestStore } from '../testing/app-store-test-fixture' -import { makeTab, makeLayout } from './terminal-tab.test-fixture' -import { createStoreSessionMockApi } from '../testing/app-store-session-api.test-fixture' -import { makeBrowserTab } from '../slices/browser/browser-tab.test-fixture' - -vi.mock('sonner', () => ({ toast: { info: vi.fn(), success: vi.fn(), error: vi.fn() } })) - -vi.mock('@/lib/agent-status/aggregation/worktree-agent-status', async (importOriginal) => { - const actual = await importOriginal() - return { - ...actual, - detectAgentStatusFromTitle: vi.fn().mockReturnValue(null) - } -}) - -createStoreSessionMockApi() - -function makeDetectedWorktreeResult( - repoId: string, - worktrees: Worktree[], - authoritative = true -): DetectedWorktreeListResult { - return { - repoId, - authoritative, - source: authoritative ? 'git' : 'metadata-fallback', - worktrees: worktrees.map((worktree) => ({ - ...worktree, - ownership: 'orca-managed', - selectedCheckout: false, - visible: true - })) - } -} - -describe('hydrateWorkspaceSession', () => { - beforeEach(() => { - vi.clearAllMocks() - }) - - it('preserves tabs for a known repo after a non-authoritative worktree fetch', () => { - // Why (#1158): metadata fallback means the runtime did not prove deletion. - const store = createTestStore() - const stalledWt = 'repo1::/path/wt1' - - store.setState({ - repos: [ - { id: 'repo1', path: '/repo1', displayName: 'Repo 1', badgeColor: '#000', addedAt: 0 } - ], - worktreesByRepo: { repo1: [] }, - detectedWorktreesByRepo: { - repo1: makeDetectedWorktreeResult('repo1', [], false) - } - }) - - store.getState().hydrateWorkspaceSession({ - activeRepoId: 'repo1', - activeWorktreeId: stalledWt, - activeTabId: 'tab1', - tabsByWorktree: { - [stalledWt]: [makeTab({ id: 'tab1', worktreeId: stalledWt })] - }, - terminalLayoutsByTabId: { - tab1: makeLayout() - } - }) - - const s = store.getState() - expect(s.tabsByWorktree[stalledWt]).toHaveLength(1) - expect(s.terminalLayoutsByTabId['tab1']).toBeDefined() - expect(s.activeWorktreeId).toBe(stalledWt) - expect(s.activeTabId).toBe('tab1') - }) - - it('drops tabs when an authoritative scan reports no matching worktrees', () => { - // Why: once git answers authoritatively, an empty repo list means deleted local worktrees, not a startup race. - const store = createTestStore() - const staleWt = 'repo1::/path/deleted' - - store.setState({ - repos: [ - { id: 'repo1', path: '/repo1', displayName: 'Repo 1', badgeColor: '#000', addedAt: 0 } - ], - worktreesByRepo: { repo1: [] }, - detectedWorktreesByRepo: { - repo1: makeDetectedWorktreeResult('repo1', []) - } - }) - - store.getState().hydrateWorkspaceSession({ - activeRepoId: 'repo1', - activeWorktreeId: staleWt, - activeTabId: 'tab-stale', - tabsByWorktree: { - [staleWt]: [makeTab({ id: 'tab-stale', worktreeId: staleWt })] - }, - terminalLayoutsByTabId: { - 'tab-stale': makeLayout() - } - }) - - const s = store.getState() - expect(s.tabsByWorktree[staleWt]).toBeUndefined() - expect(s.terminalLayoutsByTabId['tab-stale']).toBeUndefined() - expect(s.activeWorktreeId).toBeNull() - expect(s.activeTabId).toBeNull() - }) -}) - -describe('restored folder workspace hydration', () => { - beforeEach(() => { - vi.clearAllMocks() - }) - - it('keeps runtime folder workspace tabs, files, and browsers before remote catalogs load', () => { - const store = createTestStore() - const folderKey = folderWorkspaceKey('remote-folder') - const groupId = 'group-folder' - const editorFileId = '/srv/app/src/App.tsx' - const session = { - ...getDefaultWorkspaceSession(), - activeWorkspaceKey: folderKey, - activeWorktreeId: folderKey, - activeTabId: 'terminal-folder', - tabsByWorktree: { - [folderKey]: [makeTab({ id: 'terminal-folder', worktreeId: folderKey })] - }, - openFilesByWorktree: { - [folderKey]: [ - { - filePath: editorFileId, - relativePath: 'src/App.tsx', - worktreeId: folderKey, - language: 'typescript' - } - ] - }, - activeFileIdByWorktree: { [folderKey]: editorFileId }, - browserTabsByWorktree: { - [folderKey]: [ - makeBrowserTab({ - id: 'browser-folder', - worktreeId: folderKey, - url: 'https://example.com' - }) - ] - }, - browserPagesByWorkspace: { - 'browser-folder': [ - { - id: 'browser-page-folder', - workspaceId: 'browser-folder', - worktreeId: folderKey, - url: 'https://example.com', - title: 'Example', - loading: false, - faviconUrl: null, - canGoBack: false, - canGoForward: false, - loadError: null, - createdAt: 1 - } - ] - }, - activeBrowserTabIdByWorktree: { [folderKey]: 'browser-folder' }, - activeTabTypeByWorktree: { [folderKey]: 'browser' as const }, - unifiedTabs: { - [folderKey]: [ - { - id: 'terminal-folder', - entityId: 'terminal-folder', - groupId, - worktreeId: folderKey, - contentType: 'terminal' as const, - label: 'Terminal', - customLabel: null, - color: null, - sortOrder: 0, - createdAt: 1 - }, - { - id: editorFileId, - entityId: editorFileId, - groupId, - worktreeId: folderKey, - contentType: 'editor' as const, - label: 'App.tsx', - customLabel: null, - color: null, - sortOrder: 1, - createdAt: 2 - }, - { - id: 'browser-folder', - entityId: 'browser-folder', - groupId, - worktreeId: folderKey, - contentType: 'browser' as const, - label: 'Example', - customLabel: null, - color: null, - sortOrder: 2, - createdAt: 3 - } - ] - }, - tabGroups: { - [folderKey]: [ - { - id: groupId, - worktreeId: folderKey, - activeTabId: 'browser-folder', - tabOrder: ['terminal-folder', editorFileId, 'browser-folder'], - recentTabIds: ['terminal-folder', editorFileId, 'browser-folder'] - } - ] - }, - activeGroupIdByWorktree: { [folderKey]: groupId } - } - const options = { - additionalValidWorkspaceKeys: [folderKey], - runtimeHostIdByWorkspaceSessionKey: { [folderKey]: 'runtime:env-1' as const } - } - - store.getState().hydrateWorkspaceSession(session, options) - store.getState().hydrateTabsSession(session, options) - store.getState().hydrateEditorSession(session, options) - store.getState().hydrateBrowserSession(session, options) - - const state = store.getState() - expect(state.activeWorktreeId).toBe(folderKey) - expect(state.activeWorkspaceKey).toBe(folderKey) - expect(state.tabsByWorktree[folderKey]?.map((tab) => tab.id)).toEqual(['terminal-folder']) - expect(state.unifiedTabsByWorktree[folderKey]?.map((tab) => tab.id)).toEqual([ - 'terminal-folder', - editorFileId, - 'browser-folder' - ]) - expect(state.openFiles.map((file) => file.worktreeId)).toEqual([folderKey]) - expect(state.activeFileIdByWorktree[folderKey]).toBe(editorFileId) - expect(state.browserTabsByWorktree[folderKey]?.map((tab) => tab.id)).toEqual(['browser-folder']) - expect(state.browserPagesByWorkspace['browser-folder']?.[0]?.worktreeId).toBe(folderKey) - expect(state.activeTabTypeByWorktree[folderKey]).toBe('browser') - }) -}) diff --git a/src/renderer/src/store/terminals/terminal-pane-detach-agent-retention.test.ts b/src/renderer/src/store/terminals/terminal-pane-detach-agent-retention.test.ts deleted file mode 100644 index 51848ea0fbb..00000000000 --- a/src/renderer/src/store/terminals/terminal-pane-detach-agent-retention.test.ts +++ /dev/null @@ -1,240 +0,0 @@ -import { afterEach, beforeEach, describe, expect, it } from 'vitest' -import { makePaneKey } from '../../../../shared/workspace-layout/stable-pane-id' -import { collectRetainedAgentsOnDisappear } from '@/components/dashboard/useRetainedAgents' -import { - resetAgentPaneAuthorityAliasesForTests, - resolveAgentPaneAuthorityKey -} from '../slices/agent-status/agent-pane-authority' -import { createTestStore, seedStore } from '../testing/app-store-test-fixture' -import { makeTab } from './terminal-tab.test-fixture' -import { makeWorktree } from '../slices/worktrees/worktree.test-fixture' -import type { AgentStatusEntry } from '../../../../shared/agent-status/contracts/agent-status-types' -import type { TerminalTab } from '../../../../shared/workspace-layout/terminal-tab-types' -import type { DashboardAgentRow } from '@/components/dashboard/useDashboardData' - -const WORKTREE_ID = 'repo::/repo/worktree' -const SOURCE_TAB = 'tab-source' -const TARGET_TAB = 'tab-target' -const FINAL_TAB = 'tab-final' -const LEAF = '11111111-1111-4111-8111-111111111111' -const SIBLING_LEAF = '22222222-2222-4222-8222-222222222222' -const SOURCE_PANE_KEY = makePaneKey(SOURCE_TAB, LEAF) -const TARGET_PANE_KEY = makePaneKey(TARGET_TAB, LEAF) -const FINAL_PANE_KEY = makePaneKey(FINAL_TAB, LEAF) - -beforeEach(() => { - resetAgentPaneAuthorityAliasesForTests() -}) - -afterEach(() => { - resetAgentPaneAuthorityAliasesForTests() -}) - -function makeDoneEntry(args: { - paneKey: string - tabId: string - startedAt: number - agentType: AgentStatusEntry['agentType'] - terminalHandle?: string - providerSession?: AgentStatusEntry['providerSession'] -}): AgentStatusEntry { - return { - state: 'done', - prompt: 'Fix it', - updatedAt: args.startedAt, - stateStartedAt: args.startedAt, - paneKey: args.paneKey, - tabId: args.tabId, - terminalTitle: 'Claude', - stateHistory: [], - agentType: args.agentType, - terminalHandle: args.terminalHandle, - providerSession: args.providerSession, - interrupted: false - } -} - -function makeRow( - paneKey: string, - tabId: string, - startedAt = 100, - overrides?: { - agentType?: AgentStatusEntry['agentType'] - terminalHandle?: string - providerSession?: AgentStatusEntry['providerSession'] - } -): DashboardAgentRow { - const agentType = overrides?.agentType ?? 'claude' - return { - paneKey, - entry: makeDoneEntry({ - paneKey, - tabId, - startedAt, - agentType, - terminalHandle: overrides?.terminalHandle, - providerSession: overrides?.providerSession - }), - tab: makeTab({ id: tabId, worktreeId: WORKTREE_ID }), - agentType, - state: 'done', - startedAt - } -} - -function createDetachStore() { - const store = createTestStore() - seedStore(store, { - worktreesByRepo: { - repo: [makeWorktree({ id: WORKTREE_ID, repoId: 'repo', path: '/repo/worktree' })] - }, - tabsByWorktree: { - [WORKTREE_ID]: [ - makeTab({ id: SOURCE_TAB, worktreeId: WORKTREE_ID, ptyId: 'pty-a' }), - makeTab({ id: TARGET_TAB, worktreeId: WORKTREE_ID, ptyId: null }), - makeTab({ id: FINAL_TAB, worktreeId: WORKTREE_ID, ptyId: null }) - ] - }, - ptyIdsByTabId: { - [SOURCE_TAB]: ['pty-a', 'pty-sibling'], - [TARGET_TAB]: [], - [FINAL_TAB]: [] - } - }) - return store -} - -function detach( - store: ReturnType, - sourceTabId: string, - targetTabId: string -) { - store.getState().syncPaneDetachPtyOwnership({ - detachedLeafId: LEAF, - detachedPtyId: 'pty-a', - sourceLayout: { - root: { type: 'leaf', leafId: SIBLING_LEAF }, - activeLeafId: SIBLING_LEAF, - expandedLeafId: null, - ptyIdsByLeafId: { [SIBLING_LEAF]: 'pty-sibling' } - }, - sourceTabId, - targetTabId - }) -} - -function collect(args: { - currentPaneKey?: string - currentTabId?: string - currentStartedAt?: number - currentAgentType?: AgentStatusEntry['agentType'] - terminalHandle?: string - previousProviderSession?: AgentStatusEntry['providerSession'] - currentProviderSession?: AgentStatusEntry['providerSession'] - retiredPaneKeys?: Record - tabIndex?: Map -}) { - return collectRetainedAgentsOnDisappear({ - previousAgents: new Map([ - [ - SOURCE_PANE_KEY, - { - row: makeRow(SOURCE_PANE_KEY, SOURCE_TAB, 100, { - terminalHandle: args.terminalHandle, - providerSession: args.previousProviderSession - }), - worktreeId: WORKTREE_ID - } - ] - ]), - currentAgents: - args.currentPaneKey && args.currentTabId - ? new Map([ - [ - args.currentPaneKey, - { - row: makeRow(args.currentPaneKey, args.currentTabId, args.currentStartedAt, { - agentType: args.currentAgentType, - terminalHandle: args.terminalHandle, - providerSession: args.currentProviderSession - }), - worktreeId: WORKTREE_ID - } - ] - ]) - : new Map(), - retainedAgentsByPaneKey: {}, - retentionSuppressedPaneKeys: {}, - recentlyClosedAgentStatusTabIds: {}, - recentlyRetiredAgentStatusPaneKeys: args.retiredPaneKeys ?? {}, - tabIndex: args.tabIndex - }) -} - -describe('detach completed split pane → sidebar retention', () => { - it('recognizes the same run after chained transfers', () => { - const store = createDetachStore() - store.getState().setAgentStatus(SOURCE_PANE_KEY, { - state: 'done', - prompt: 'Fix it', - agentType: 'claude' - }) - - detach(store, SOURCE_TAB, TARGET_TAB) - detach(store, TARGET_TAB, FINAL_TAB) - - expect(collect({ currentPaneKey: FINAL_PANE_KEY, currentTabId: FINAL_TAB }).toRetain).toEqual( - [] - ) - // Why: every intermediate key must resolve to the CURRENT owner, not the previous hop. - expect(resolveAgentPaneAuthorityKey(SOURCE_PANE_KEY)).toBe(FINAL_PANE_KEY) - expect(resolveAgentPaneAuthorityKey(TARGET_PANE_KEY)).toBe(FINAL_PANE_KEY) - }) - - it('does not resurrect the source when the transferred owner immediately closes', () => { - const store = createDetachStore() - store.getState().setAgentStatus(SOURCE_PANE_KEY, { - state: 'done', - prompt: 'Fix it', - agentType: 'claude' - }) - - detach(store, SOURCE_TAB, TARGET_TAB) - store.getState().dropAgentStatusByTabPrefix(TARGET_TAB) - const retiredPaneKeys = store.getState().recentlyRetiredAgentStatusPaneKeys - - expect(retiredPaneKeys[SOURCE_PANE_KEY]).toBe(true) - expect(retiredPaneKeys[TARGET_PANE_KEY]).toBe(true) - expect(collect({ retiredPaneKeys }).toRetain).toEqual([]) - }) - - it('retains at the transferred surface when the destination pty exits before the next sync', () => { - const store = createDetachStore() - store.getState().setAgentStatus(SOURCE_PANE_KEY, { - state: 'done', - prompt: 'Fix it', - agentType: 'claude' - }) - - detach(store, SOURCE_TAB, TARGET_TAB) - store.getState().removeAgentStatus(TARGET_PANE_KEY) - - // Why: the retention effect coalesced the move and the exit into one pass, so it - // only ever sees the source key vanish. - const destinationTab = makeTab({ - id: TARGET_TAB, - worktreeId: WORKTREE_ID, - title: 'Destination', - launchAgent: 'codex' - }) - const retained = collect({ - tabIndex: new Map([[TARGET_TAB, { tab: destinationTab }]]) - }).toRetain - expect(retained).toHaveLength(1) - expect(retained[0]?.entry.paneKey).toBe(TARGET_PANE_KEY) - expect(retained[0]?.entry.tabId).toBe(TARGET_TAB) - // Retain destination tab identity; sidebar agent resolution reads its title and launchAgent. - expect(retained[0]?.tab).toBe(destinationTab) - expect(retained[0]?.tab.title).toBe('Destination') - }) -}) From 3d9be7c2696debfc6133e3000a558f9afbdb45d3 Mon Sep 17 00:00:00 2001 From: Neil <4138956+nwparker@users.noreply.github.com> Date: Tue, 10 Nov 2026 16:58:38 -0800 Subject: [PATCH 12/21] relay: an edge fence refused as a same-epoch conflict goes once more one epoch above (#40946) At 1% the edge answered fences 'conflict': it holds the fence's own epoch for another placement (a move onto the edge whose push landed though the director counted it failed, then a placement elsewhere minted the same epoch). Retries go to the host's current row at that same epoch, so they kept failing until the host moved again; meanwhile the edge object claims the host and its phones get HOST_OFFLINE instead of re-resolving. No later move can produce a same-epoch conflict (it would answer stale), so the fence now goes once more one epoch above, which lands; a later move onto the edge re-mints above it. Stale answers are never bumped. --- cloud/apps/relay/src/edge-relay-fence.test.ts | 37 +++++++++++++++++++ cloud/apps/relay/src/edge-relay-fence.ts | 25 ++++++++++++- 2 files changed, 60 insertions(+), 2 deletions(-) diff --git a/cloud/apps/relay/src/edge-relay-fence.test.ts b/cloud/apps/relay/src/edge-relay-fence.test.ts index abdef6a01ba..bfee89d38c3 100644 --- a/cloud/apps/relay/src/edge-relay-fence.test.ts +++ b/cloud/apps/relay/src/edge-relay-fence.test.ts @@ -116,6 +116,43 @@ describe('relay edge fence after a host leaves', () => { }) }) +describe('relay edge fence over a same-epoch conflict', () => { + it('fences once more one epoch above a conflict at its epoch, and never bumps over a stale answer', async () => { + vi.spyOn(console, 'error').mockImplementation(() => {}) + const warn = vi.spyOn(console, 'warn').mockImplementation(() => {}) + const h = await harness() + h.edge.answers = ['conflict', 'applied'] + expect(await h.assignments.fenceAway({ userId: USER, relayHostId: HOST }, 'us-c1', 9)).toBe(true) + expect(h.edge.calls.map((call) => call.body.assignmentEpoch)).toEqual([9, 10]) + expect(h.edge.calls.every((call) => call.body.cellId === 'us-c1')).toBe(true) + expect(logged(warn, 'orca_relay_edge_fence_above_conflict')).toEqual([ + { event: 'orca_relay_edge_fence_above_conflict', outcome: 'applied', assignmentEpoch: 10 } + ]) + expect(await h.pending()).toEqual([]) + // Stale: the edge holds a newer epoch (the host may be back there); no bump, recorded for retry. + h.edge.calls = [] + h.edge.answers = ['stale'] + expect(await h.assignments.fenceAway({ userId: USER, relayHostId: HOST }, 'us-c1', 11)).toBe(false) + expect(h.edge.calls.map((call) => call.body.assignmentEpoch)).toEqual([11]) + }) + + it('a pending fence retried into a same-epoch conflict lands one epoch above the row', async () => { + vi.spyOn(console, 'error').mockImplementation(() => {}) + vi.spyOn(console, 'warn').mockImplementation(() => {}) + const h = await harness() + await h.placeRow('us-c1', 7) + h.edge.answers = ['down', 'down', 'down', 'down', 'down', 'down', 'down', 'down', 'down', 'down'] + expect(await h.assignments.fenceAway({ userId: USER, relayHostId: HOST }, 'us-c1', 7)).toBe(false) + expect(await h.pending()).toHaveLength(1) + h.edge.calls = [] + h.edge.answers = ['conflict', 'applied'] + h.clock.now += 60_000 + await h.retry() + expect(h.edge.calls.map((call) => call.body.assignmentEpoch)).toEqual([7, 8]) + expect(await h.pending()).toEqual([]) + }) +}) + describe('relay edge pending fence retries', () => { it('re-fences to the current row, not the recorded one, and clears the record once it lands', async () => { const warn = vi.spyOn(console, 'warn').mockImplementation(() => {}) diff --git a/cloud/apps/relay/src/edge-relay-fence.ts b/cloud/apps/relay/src/edge-relay-fence.ts index 3d01795ddaa..e18056b5eeb 100644 --- a/cloud/apps/relay/src/edge-relay-fence.ts +++ b/cloud/apps/relay/src/edge-relay-fence.ts @@ -41,6 +41,27 @@ export async function recordPendingFence( ) } +// One fence call. `conflict` means the edge holds this very epoch for another placement (a move onto +// the edge whose push landed though the director counted it as failed, then a placement elsewhere +// minted the same epoch). No later move can produce that (a later move has a higher epoch, which +// answers `stale`), so the fence goes once more one epoch above: it lands, and a later move onto +// the edge re-mints above it as for any edge epoch (moveToEdge). +async function assignFence( + admin: EdgeAdmin, + fence: Identity & { cellId: string; assignmentEpoch: number } +): Promise<{ outcome: string; assignmentEpoch: number }> { + const first = await admin.assign(fence) + if (first.outcome !== 'conflict' || first.assignmentEpoch !== fence.assignmentEpoch) { + return { outcome: first.outcome, assignmentEpoch: fence.assignmentEpoch } + } + const above = fence.assignmentEpoch + 1 + const second = await admin.assign({ ...fence, assignmentEpoch: above }) + console.warn( + JSON.stringify({ event: 'orca_relay_edge_fence_above_conflict', outcome: second.outcome, assignmentEpoch: above }) + ) + return { outcome: second.outcome, assignmentEpoch: above } +} + // The fence right after a host left (EdgeAssignments.fenceAway). Anything but applied/unchanged // fails: unavailable, `conflict`, or `stale` (the edge holds a newer epoch, so it still claims the // host); the failure is logged and recorded for retry. Returns whether the fence landed. @@ -55,7 +76,7 @@ export async function fenceEdge( let outcome = 'unavailable' for (let attempt = 0; attempt < attempts; attempt += 1) { try { - outcome = (await admin.assign(fence)).outcome + outcome = (await assignFence(admin, fence)).outcome } catch (error) { if (!(error instanceof EdgeAdminUnavailableError)) { throw error } outcome = error.message @@ -138,7 +159,7 @@ async function retryOne( const assignmentEpoch = integer(row, 'assignment_epoch') let outcome: string try { - outcome = (await admin.assign({ ...identity, cellId, assignmentEpoch })).outcome + outcome = (await assignFence(admin, { ...identity, cellId, assignmentEpoch })).outcome } catch (error) { if (!(error instanceof EdgeAdminUnavailableError)) { throw error } outcome = error.message From bc35323f177fcd44706c8cefaadb165e27a3b302 Mon Sep 17 00:00:00 2001 From: Brennan Benson <79079362+brennanb2025@users.noreply.github.com> Date: Tue, 10 Nov 2026 17:06:19 -0800 Subject: [PATCH 13/21] Keep Chat UI model selections from snapping back to older replies (#40906) * fix(native-chat): keep a model pick shown instead of an older answer's model A saved model pick on a live structured Claude chat snapped back: the host's options answer kept returning the model that answered before the pick, and the picker shows a reported model over the selection (for provider fallback). A pick now retires the child's earlier model reports until a turn carries that pick, as the common pattern does with a fallback report once the user picks again. The Claude adapter ignores old-execution frames until a turn applies the picked value; the host drops a report captured by an options read when a pick or a child change lands before the read answers. Picks still reach Claude at the next turn. * test(native-chat): cover the confirmed-init report under read freshness * test(native-chat): cover picks during recovery and legacy no-op mutations --- ...claude-context-usage-unloaded-turn.test.ts | 3 +- ...laude-structured-execution-observations.ts | 12 +- .../claude-structured-response-model.test.ts | 47 +++++ .../claude-structured-session-adapter.ts | 7 + .../claude-structured-session-state.ts | 2 + .../claude-structured-message-options.ts | 4 + .../claude-structured-session-options.ts | 5 +- ...ent-session-option-refresh.test-fixture.ts | 3 +- ...sion-options-read-report-freshness.test.ts | 182 ++++++++++++++++++ ...structured-agent-session-host-mutations.ts | 6 +- .../structured-agent-session-option-intent.ts | 35 ++++ .../structured-agent-session-options-read.ts | 28 ++- ...structured-agent-session-adapter-router.ts | 2 + .../structured-agent-session-adapter.ts | 2 + 14 files changed, 321 insertions(+), 17 deletions(-) create mode 100644 src/main/native-chat/agent-session-wire/host/testing/structured-agent-session-options-read-report-freshness.test.ts diff --git a/src/main/claude/claude-context-usage-unloaded-turn.test.ts b/src/main/claude/claude-context-usage-unloaded-turn.test.ts index bd2d8c1edd4..fcfcffed4aa 100644 --- a/src/main/claude/claude-context-usage-unloaded-turn.test.ts +++ b/src/main/claude/claude-context-usage-unloaded-turn.test.ts @@ -108,7 +108,8 @@ function readOptions(journal: AgentSessionJournal, agents: StructuredAgentRegist }, serialize: (_sessionId: string, task: () => Promise) => task(), openConversation: async () => running, - conversation: async () => running + conversation: async () => running, + sessions: new Map([[SESSION, running]]) } as unknown as StructuredAgentSessionMutationContext return readStructuredAgentSessionOptions(context, SESSION, { readsWithoutModel: true }) } diff --git a/src/main/claude/structured/claude-structured-execution-observations.ts b/src/main/claude/structured/claude-structured-execution-observations.ts index bd27887fdfd..60f7a34ecaf 100644 --- a/src/main/claude/structured/claude-structured-execution-observations.ts +++ b/src/main/claude/structured/claude-structured-execution-observations.ts @@ -41,11 +41,19 @@ export function beginClaudeOptionMutation(session: ClaudeSession, key: string): return sequence } +/** Until a turn carries the pick, the running process's answers name the model it replaced. */ +export function noteClaudeModelPick(session: ClaudeSession, model: string): void { + session.modelPickAwaitingTurn = model + delete session.reportedOptions.reportedModel +} + /** An init names execution; a control answer only acknowledges its input. */ export function observeClaudeExecutionModel(session: ClaudeSession, model: string): boolean { const previous = claudeExecutionObservations(session).model const previousResponse = session.reportedOptions.reportedModel - session.reportedOptions.reportedModel = model + if (!session.modelPickAwaitingTurn) { + session.reportedOptions.reportedModel = model + } if (previous !== model) { delete session.appliedOptions } @@ -93,7 +101,7 @@ function observeClaudeResponseModel( : null, responseModel: response }) ?? response - if (model === session.reportedOptions.reportedModel) { + if (session.modelPickAwaitingTurn || model === session.reportedOptions.reportedModel) { return false } session.reportedOptions.reportedModel = model diff --git a/src/main/claude/structured/claude-structured-response-model.test.ts b/src/main/claude/structured/claude-structured-response-model.test.ts index e37af00c246..4405ad02c96 100644 --- a/src/main/claude/structured/claude-structured-response-model.test.ts +++ b/src/main/claude/structured/claude-structured-response-model.test.ts @@ -152,3 +152,50 @@ it('keeps a model picked during a turn over that turn’s later replies until a await adapter.closeAll() } }) + +it('retires the answered model on a saved pick until a turn carries that pick', async () => { + const claude = fakeClaude({ initModel: 'claude-opus-5-5' }) + const adapter = await acquired(claude) + const connection = claude.connections[0]! + const read = async () => (await adapter.readOptions({ sessionId: 'session-1', fence: 7 })).current + const answer = (uuid: string, model: string) => + claudeFrame(connection, { + type: 'assistant', + uuid, + message: { role: 'assistant', model, content: [] } + }) + const turn = async (clientMessageId: string, model: string) => { + claudeFrame(connection, { type: 'system', subtype: 'session_state_changed', state: 'idle' }) + await adapter.dispatch({ + sessionId: 'session-1', + clientMessageId, + body: USER_MESSAGE, + executionOptions: { model }, + fence: 7 + }) + } + try { + claudeFrame(connection, { type: 'system', subtype: 'init', model: 'claude-opus-5-5' }) + answer('fallback', 'claude-opus-4-8') + expect((await read()).reportedModel).toBe('claude-opus-4-8') + + adapter.noteOptionPick({ sessionId: 'session-1', fence: 7, key: 'model', value: 'sonnet' }) + answer('same-turn', 'claude-opus-4-8') + expect((await read()).reportedModel).toBeUndefined() + expect((await read()).confirmed ?? []).not.toContain('model') + + // A message accepted before the pick runs the model it captured. + await turn('queued-before-pick', 'opus') + claudeFrame(connection, { type: 'system', subtype: 'init', model: 'claude-opus-5-5' }) + answer('queued-answer', 'claude-opus-5-5') + expect((await read()).reportedModel).toBeUndefined() + + await turn('carries-pick', 'sonnet') + expect(connection.calls).toContainEqual({ subtype: 'set_model', params: { model: 'sonnet' } }) + claudeFrame(connection, { type: 'system', subtype: 'init', model: 'claude-sonnet-5-5' }) + answer('next-fallback', 'claude-opus-4-8') + expect((await read()).reportedModel).toBe('claude-opus-4-8') + } finally { + await adapter.closeAll() + } +}) diff --git a/src/main/claude/structured/claude-structured-session-adapter.ts b/src/main/claude/structured/claude-structured-session-adapter.ts index 53d4d7a7df8..e9369cceb69 100644 --- a/src/main/claude/structured/claude-structured-session-adapter.ts +++ b/src/main/claude/structured/claude-structured-session-adapter.ts @@ -18,6 +18,7 @@ import { supportsClaudeStructuredLocation } from './launch/claude-structured-loc import * as claudeOptions from './launch/claude-structured-options' import { dispatchClaudeMessage } from './launch/claude-structured-message-options' import { readClaudeStructuredSessionOptions } from './launch/claude-structured-session-options' +import { noteClaudeModelPick } from './claude-structured-execution-observations' import { ClaudeAcquisitionRegistry, type ClaudeSession, @@ -231,6 +232,12 @@ export class ClaudeStructuredSessionAdapter implements StructuredAgentSessionAda deps: this.deps, emit: (session, event) => this.emit(session, event) }) + noteOptionPick: NonNullable = (input) => { + const session = this.sessions.get(input.sessionId) + if (session?.fence === input.fence && input.key === 'model' && input.value) { + noteClaudeModelPick(session, input.value) + } + } readOptions = (input: { sessionId: string; fence: number }) => readClaudeStructuredSessionOptions(this.session(input.sessionId)) releaseAcquisition = (input: ClaudeReleaseRequest): Promise => { diff --git a/src/main/claude/structured/claude-structured-session-state.ts b/src/main/claude/structured/claude-structured-session-state.ts index 55959e4900f..130bc0f0441 100644 --- a/src/main/claude/structured/claude-structured-session-state.ts +++ b/src/main/claude/structured/claude-structured-session-state.ts @@ -135,6 +135,8 @@ export type ClaudeSession = { fastModePerSessionOptIn?: boolean /** Mutation sequence of the last model report; newer writes outrank it. */ reportedModelMutation: number + /** A saved model pick no turn of this process has carried; its reports predate the pick. */ + modelPickAwaitingTurn?: string /** `dispatchSequence` at the last model write; that turn's replies predate the pick. */ modelMutationDispatch?: number /** Options whose recorded value the provider reported, not merely accepted. */ diff --git a/src/main/claude/structured/launch/claude-structured-message-options.ts b/src/main/claude/structured/launch/claude-structured-message-options.ts index 8463df786cc..abd34fe3d5b 100644 --- a/src/main/claude/structured/launch/claude-structured-message-options.ts +++ b/src/main/claude/structured/launch/claude-structured-message-options.ts @@ -30,6 +30,10 @@ export async function applyClaudeMessageOptions( ): Promise { throwIfSignalAborted(signal) const model = options.model + // Only a turn carrying the newest pick reports on it; earlier captured sends run the replaced model. + if (model === session.modelPickAwaitingTurn) { + delete session.modelPickAwaitingTurn + } if (model && session.appliedInputOptions.get('model') !== model) { await waitForPromiseWithSignal( setClaudeStructuredOption(session, { key: 'model', value: model }, timeoutMs), diff --git a/src/main/claude/structured/launch/claude-structured-session-options.ts b/src/main/claude/structured/launch/claude-structured-session-options.ts index 000c85ec47e..b35843bd7c6 100644 --- a/src/main/claude/structured/launch/claude-structured-session-options.ts +++ b/src/main/claude/structured/launch/claude-structured-session-options.ts @@ -213,6 +213,7 @@ export function claudeStructuredSessionOptionsFrom( const discovered = listedModels(catalog ? { models: catalog } : null) const listed = discovered.length > 0 ? discovered : seedModels() const current = readClaudeCurrentModel(session) + const reportsPick = !session.modelPickAwaitingTurn const model = current.id === 'default' ? undefined : current.id const models = structuredAgentSessionOptionModels(listed, model, (row) => ({ ...row, @@ -230,7 +231,7 @@ export function claudeStructuredSessionOptionsFrom( session.reportedOptions.fastMode ?? (session.fastModeState === undefined ? undefined : session.fastModeState !== 'off') const confirmed = [ - ...(current.confirmed ? ['model'] : []), + ...(current.confirmed && reportsPick ? ['model'] : []), ...(effort && session.confirmedOptions.has('effort') ? ['effort'] : []), ...(fastMode !== undefined && (session.confirmedOptions.has('fastMode') || !session.options.has('fastMode')) @@ -243,7 +244,7 @@ export function claudeStructuredSessionOptionsFrom( ...(changesLive ? {} : { effortFixed: true as const }), current: { model, - reportedModel: session.reportedOptions.reportedModel, + reportedModel: reportsPick ? session.reportedOptions.reportedModel : undefined, reportedPermissionMode: session.reportedOptions.reportedPermissionMode, ...(effort ? { effort } : {}), ...(fastMode !== undefined ? { fastMode } : {}), diff --git a/src/main/native-chat/agent-session-wire/host/testing/structured-agent-session-option-refresh.test-fixture.ts b/src/main/native-chat/agent-session-wire/host/testing/structured-agent-session-option-refresh.test-fixture.ts index bc258adeaed..7fb76b5ae44 100644 --- a/src/main/native-chat/agent-session-wire/host/testing/structured-agent-session-option-refresh.test-fixture.ts +++ b/src/main/native-chat/agent-session-wire/host/testing/structured-agent-session-option-refresh.test-fixture.ts @@ -106,7 +106,8 @@ export function piOptionRefreshHost(sessionId = 'session-1') { }, serialize: (_sessionId: string, task: () => Promise) => task(), openConversation: async () => conversation, - conversation: async () => conversation + conversation: async () => conversation, + sessions: new Map([[sessionId, conversation]]) } as unknown as StructuredAgentSessionMutationContext return readStructuredAgentSessionOptions(context, sessionId, { readsWithoutModel: true }) }, diff --git a/src/main/native-chat/agent-session-wire/host/testing/structured-agent-session-options-read-report-freshness.test.ts b/src/main/native-chat/agent-session-wire/host/testing/structured-agent-session-options-read-report-freshness.test.ts new file mode 100644 index 00000000000..cd25747e37b --- /dev/null +++ b/src/main/native-chat/agent-session-wire/host/testing/structured-agent-session-options-read-report-freshness.test.ts @@ -0,0 +1,182 @@ +import { expect, it } from 'vitest' +import type { AgentSessionModelCatalogResult } from '../../../../../shared/agent-session/agent-session-wire' +import type { StructuredAgentSessionMutationContext } from '../../operations/structured-agent-session-mutation-context' +import type { + StructuredAgentSessionHostSession, + StructuredAgentSessionProviderChild +} from '../structured-agent-session-host-types' +import { readStructuredAgentSessionOptions } from '../../options/structured-agent-session-options-read' +import { noteStructuredAgentSessionOptionPick } from '../../options/structured-agent-session-option-intent' +import { + hostTestCatalogReadOperation, + hostTestOptionsReadStore, + hostTestRestingOptionsJournal, + hostTestRestingOptionsRecord, + HOST_TEST_NOW as NOW, + HOST_TEST_SESSION as SESSION, + hostTestOperationId +} from './structured-agent-session-host-test-data' +import { + createQueuedMessageTestRig, + QUEUED_RIG_CALLER +} from './structured-agent-session-queued-message-rig.test-fixture' +import { appendUsageLimitedTurn } from './structured-agent-session-rewind-usage-limit.test-fixture' + +const RESPONSE_REPORT = { model: 'opus[1m]', reportedModel: 'claude-opus-4-8' } +// Claude's confirmed init model stands in for a report when no response named one. +const CONFIRMED_INIT = { model: 'opus[1m]', confirmed: ['model'] } + +function claudeOptionsRead(answer: Record = RESPONSE_REPORT) { + const record = { + ...hostTestRestingOptionsRecord('session-1'), + provider: 'claude', + options: { model: 'opus[1m]' } + } + // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: the reader and pick counter use only the child's identity, fence and phase. + const child = { fence: 2, phase: 'ready' } as unknown as StructuredAgentSessionProviderChild + const conversation = { + child, + params: { provider: 'claude' }, + journal: hostTestRestingOptionsJournal() + } + // Picks serialize with the live read, so they can land only after it, as the answer is assembled. + const entered = Promise.withResolvers() + const assembled = Promise.withResolvers() + const home: AgentSessionModelCatalogResult = { origin: 'probe', fetchedAt: 1, models: [] } + // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: picks and the reader read only the session's current child. + const session = { child } as unknown as StructuredAgentSessionHostSession + const sessions = new Map([['session-1', session]]) + // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: the real reader accesses only these injected ports; no provider process is launched. + const context = { + deps: { + adapter: { + readOptions: async () => ({ + models: [], + current: answer + }) + }, + agents: { definition: () => null, capabilities: () => null }, + store: hostTestOptionsReadStore(() => record), + modelCatalog: { + captureRead: hostTestCatalogReadOperation, + read: async () => home + } + }, + serialize: (_sessionId: string, task: () => Promise) => task(), + openConversation: async () => conversation, + conversation: async () => { + entered.resolve() + await assembled.promise + return conversation + }, + sessions + } as unknown as StructuredAgentSessionMutationContext + const pick = (model: string) => { + record.options.model = model + noteStructuredAgentSessionOptionPick( + { deps: { adapter: {} }, sessions }, + { sessionId: 'session-1', fence: 2 }, + { key: 'model', value: model }, + true + ) + } + /** Run `during` while the read is between its live snapshot and its answer. */ + const read = async (during: () => void = () => {}) => { + const answer = readStructuredAgentSessionOptions(context, 'session-1', { + readsWithoutModel: true + }) + await entered.promise + during() + assembled.resolve() + return (await answer).current + } + return { read, pick, sessions, session } +} + +it('keeps a live fallback report that no pick has replaced', async () => { + expect(await claudeOptionsRead(CONFIRMED_INIT).read()).toMatchObject({ + reportedModel: 'opus[1m]' + }) + expect(await claudeOptionsRead().read()).toMatchObject({ + model: 'opus[1m]', + reportedModel: 'claude-opus-4-8' + }) +}) + +it.each([ + ['picks back to the same model', ['sonnet', 'opus[1m]'], RESPONSE_REPORT], + ['re-picks the shown model', ['opus[1m]'], RESPONSE_REPORT], + ['picks back while only the init model confirms', ['sonnet', 'opus[1m]'], CONFIRMED_INIT] +])( + 'drops a report captured before the user %s while the read is in flight', + async (_, models, answer) => { + const host = claudeOptionsRead(answer) + const current = await host.read(() => models.forEach((model) => host.pick(model))) + expect(current.model).toBe('opus[1m]') + expect(current.reportedModel).toBeUndefined() + } +) + +it('drops a report from a child that ended or was replaced while the read was in flight', async () => { + for (const next of [null, { fence: 3, phase: 'ready' }]) { + const host = claudeOptionsRead() + const current = await host.read(() => { + // oxlint-disable-next-line typescript/consistent-type-assertions -- SAFETY: only child identity matters to the reader. + host.sessions.set('session-1', { + ...host.session, + child: next + } as unknown as StructuredAgentSessionHostSession) + }) + expect(current.reportedModel).toBeUndefined() + } +}) + +it.each([ + [ + 'retires the report when the user picks back to it', + QUEUED_RIG_CALLER, + ['sonnet', 'opus[1m]'], + undefined + ], + [ + 'keeps the report through an old client empty no-op', + { ...QUEUED_RIG_CALLER, readsOptionsWithoutModel: false }, + [''], + 'claude-opus-4-8' + ] +] as const)( + 'after a usage-limited turn, a pick made while recovery choices load %s', + async (_, caller, models, reportedModel) => { + const rig = await createQueuedMessageTestRig({ agent: 'claude' }) + try { + const pick = async (value: string) => { + const fields = { key: 'model', value } + return rig.host.setOption(caller, { + envelope: rig.envelope(fields, 'agentSession.setOption', hostTestOperationId()), + ...fields + }) + } + expect(await pick('opus[1m]')).toMatchObject({ ok: true }) + rig.host.deps.adapter.readOptions = async () => ({ models: [], current: RESPONSE_REPORT }) + const journal = rig.host.collaboratorsForTests().sessions.get(SESSION)!.journal + await appendUsageLimitedTurn(journal, NOW, 'quota') + const recovering = Promise.withResolvers() + const released = Promise.withResolvers() + rig.host.deps.resolveLaunchEnv = () => { + recovering.resolve() + return released.promise + } + const answer = rig.host.readOptions(SESSION) + await recovering.promise + for (const model of models) { + expect(await pick(model)).toMatchObject({ ok: true }) + } + released.resolve(undefined) + const { current } = await answer + expect(current.model).toBe('opus[1m]') + expect(current.reportedModel).toBe(reportedModel) + } finally { + await rig.dispose() + } + } +) diff --git a/src/main/native-chat/agent-session-wire/operations/structured-agent-session-host-mutations.ts b/src/main/native-chat/agent-session-wire/operations/structured-agent-session-host-mutations.ts index e05d409093c..664ea0022cc 100644 --- a/src/main/native-chat/agent-session-wire/operations/structured-agent-session-host-mutations.ts +++ b/src/main/native-chat/agent-session-wire/operations/structured-agent-session-host-mutations.ts @@ -55,7 +55,10 @@ import { openWithReadyStructuredAgentSessionChild } from './structured-agent-ses import type { StructuredAgentSessionCaller } from '../host/structured-agent-session-host-types' import '../options/structured-agent-session-options-read' -import { recordStructuredAgentSessionOptionIntent } from '../options/structured-agent-session-option-intent' +import { + noteStructuredAgentSessionOptionPick, + recordStructuredAgentSessionOptionIntent +} from '../options/structured-agent-session-option-intent' import { confirmIdleStructuredAgentSessionOption } from '../options/structured-agent-session-idle-option' export function sendStructuredAgentSessionTurn( @@ -265,6 +268,7 @@ export async function setStructuredAgentSessionOption( readsWithoutModel: caller.readsOptionsWithoutModel !== false }) if (picked.ok) { + noteStructuredAgentSessionOptionPick(context, ctx, params, !noChoice) confirmIdleStructuredAgentSessionOption(context, ctx, params) } return picked diff --git a/src/main/native-chat/agent-session-wire/options/structured-agent-session-option-intent.ts b/src/main/native-chat/agent-session-wire/options/structured-agent-session-option-intent.ts index 2de64669af4..f835e253b20 100644 --- a/src/main/native-chat/agent-session-wire/options/structured-agent-session-option-intent.ts +++ b/src/main/native-chat/agent-session-wire/options/structured-agent-session-option-intent.ts @@ -19,6 +19,10 @@ import { } from './structured-agent-session-pick-reply' import type { StructuredAgentSessionAdapter } from '../providers/structured-agent-session-adapter' import { pickAgentSessionOption } from '../../../runtime/agent-session/records/agent-session-record-options' +import type { + StructuredAgentSessionHostSession, + StructuredAgentSessionProviderChild +} from '../host/structured-agent-session-host-types' function acceptsOptionPick( agents: Pick, @@ -103,3 +107,34 @@ export async function recordStructuredAgentSessionOptionIntent( } } } + +const modelPicks = new WeakMap() + +/** Counts saved model picks per child, so a read can tell its captured report predates one. */ +export function structuredAgentSessionModelPicks( + child: StructuredAgentSessionProviderChild +): number { + return modelPicks.get(child) ?? 0 +} + +/** A saved pick retires the current child's earlier reports, including ones a read already captured. */ +export function noteStructuredAgentSessionOptionPick( + context: { + deps: { adapter: Pick } + sessions: ReadonlyMap + }, + ctx: Pick, + input: { key: string; value: string }, + saved: boolean +): void { + const { sessionId, fence } = ctx + // An old client's empty model clears nothing, so it retires no report. + if (!saved) { + return + } + const child = context.sessions.get(sessionId)?.child + if (input.key === 'model' && child?.fence === fence) { + modelPicks.set(child, structuredAgentSessionModelPicks(child) + 1) + } + context.deps.adapter.noteOptionPick?.({ sessionId, fence, ...input }) +} diff --git a/src/main/native-chat/agent-session-wire/options/structured-agent-session-options-read.ts b/src/main/native-chat/agent-session-wire/options/structured-agent-session-options-read.ts index e2c14973552..76918d8fe34 100644 --- a/src/main/native-chat/agent-session-wire/options/structured-agent-session-options-read.ts +++ b/src/main/native-chat/agent-session-wire/options/structured-agent-session-options-read.ts @@ -30,12 +30,13 @@ import { structuredAgentSessionImmediatePermission } from './structured-agent-se import type { StructuredAgentSessionMutationContext } from '../operations/structured-agent-session-mutation-context' import { readStructuredSessionRecovery } from '../recovery/structured-agent-session-recovery-read' import { fillUnknownModelEfforts } from '../../../../shared/agent-session/options/agent-session-model-efforts' +import { structuredAgentSessionModelPicks } from './structured-agent-session-option-intent' /** A live child's own answer, or the answer at rest; the goal, usage and rewind either way. */ export async function readStructuredAgentSessionOptions( context: Pick< StructuredAgentSessionMutationContext, - 'deps' | 'serialize' | 'openConversation' | 'conversation' + 'deps' | 'serialize' | 'openConversation' | 'conversation' | 'sessions' > & Partial>, sessionId: string, @@ -71,6 +72,7 @@ export async function readStructuredAgentSessionOptions( child && child.phase !== 'starting' ? await adapter.readOptions?.({ sessionId, fence: child.fence }) : undefined + const modelPicks = child ? structuredAgentSessionModelPicks(child) : 0 const scope = () => structuredAgentSessionCatalogScopeOwned( context.deps, @@ -93,7 +95,8 @@ export async function readStructuredAgentSessionOptions( records, session, child, - options: observed ?? null + options: observed ?? null, + modelPicks } }) let started = await readCurrent(true) @@ -132,6 +135,7 @@ export async function readStructuredAgentSessionOptions( const floor = session.journal.context.floor() const contextFloor = floor ? { contextFloor: floor } : {} const restingOptions = await atRest() + const recovery = await readStructuredSessionRecovery(session, context, sessionId) if (live && usesModelSpecificAcpEfforts(agent)) { options = { ...options, @@ -159,12 +163,17 @@ export async function readStructuredAgentSessionOptions( const model = saved.model ?? (options.current.model || selected.current.model || displayDefault) const projected = live?.projectCurrent?.({ ...saved, ...(model ? { model } : {}) }) const reportedEffort = projected?.confirmed?.includes('effort') ? projected.effort : undefined - const reportedModel = live - ? (options.current.reportedModel ?? - (agent === 'claude' && options.current.confirmed?.includes('model') - ? options.current.model - : undefined)) - : undefined + // After every await: a pick or a replaced child since the live read retires its report. + const reportedModel = + live && + started.kind === 'live' && + context.sessions.get(sessionId)?.child === started.child && + started.modelPicks === structuredAgentSessionModelPicks(started.child) + ? (options.current.reportedModel ?? + (agent === 'claude' && options.current.confirmed?.includes('model') + ? options.current.model + : undefined)) + : undefined const current = { ...selected.current, ...options.current, @@ -236,7 +245,6 @@ export async function readStructuredAgentSessionOptions( current.serviceTier = projected.options.serviceTier } } - const recovery = await readStructuredSessionRecovery(session, context, sessionId) return { ...options, ...recovery, @@ -258,7 +266,7 @@ export async function readStructuredAgentSessionOptions( current: { ...current, acceptedInputSources, - ...(reportedModel ? { reportedModel } : {}), + reportedModel, ...(reportedEffort !== undefined ? { reportedEffort } : {}), defaulted, confirmed, diff --git a/src/main/native-chat/agent-session-wire/providers/structured-agent-session-adapter-router.ts b/src/main/native-chat/agent-session-wire/providers/structured-agent-session-adapter-router.ts index 492ff9bc30c..978847e8632 100644 --- a/src/main/native-chat/agent-session-wire/providers/structured-agent-session-adapter-router.ts +++ b/src/main/native-chat/agent-session-wire/providers/structured-agent-session-adapter-router.ts @@ -204,6 +204,8 @@ export class StructuredAgentSessionAdapterRouter implements Adapter { canSetOptionWhileIdle: Adapter['canSetOptionWhileIdle'] = (input) => this.liveOwnerOrNull(input.sessionId)?.canSetOptionWhileIdle?.(input) === true + noteOptionPick: Adapter['noteOptionPick'] = (input) => + this.liveOwnerOrNull(input.sessionId)?.noteOptionPick?.(input) readAppliedInputOptions: Adapter['readAppliedInputOptions'] = (input) => this.liveOwnerOrNull(input.sessionId)?.readAppliedInputOptions?.(input) ?? null diff --git a/src/main/native-chat/agent-session-wire/providers/structured-agent-session-adapter.ts b/src/main/native-chat/agent-session-wire/providers/structured-agent-session-adapter.ts index 31a3adcbf6b..2a5269d08e9 100644 --- a/src/main/native-chat/agent-session-wire/providers/structured-agent-session-adapter.ts +++ b/src/main/native-chat/agent-session-wire/providers/structured-agent-session-adapter.ts @@ -304,6 +304,8 @@ export type StructuredAgentSessionAdapter = StructuredAgentSessionAdapterStop & ): Promise>> /** Opt in only when setOption can update idle metadata without a provider turn. */ canSetOptionWhileIdle?(input: StructuredAgentSessionSetOptionInput): boolean + /** A saved pick retires this child's earlier execution reports until a turn carries it. */ + noteOptionPick?(input: { sessionId: string; fence: number; key: string; value: string }): void /** Applied inputs for this exact child; observations never decide message admission. */ readAppliedInputOptions?(input: { sessionId: string From 2259dbe1631e5ab022af4d6d5dfa82a9a6c49a6f Mon Sep 17 00:00:00 2001 From: Kelvin Amoaba <97001695+AmoabaKelvin@users.noreply.github.com> Date: Wed, 11 Nov 2026 01:07:50 +0000 Subject: [PATCH 14/21] fix(status-bar): import window visibility interval from its new path (#40951) --- .../components/status-bar/workspace-space-issue-freshness.tsx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/renderer/src/components/status-bar/workspace-space-issue-freshness.tsx b/src/renderer/src/components/status-bar/workspace-space-issue-freshness.tsx index 004742c7f9b..97d0ece42b6 100644 --- a/src/renderer/src/components/status-bar/workspace-space-issue-freshness.tsx +++ b/src/renderer/src/components/status-bar/workspace-space-issue-freshness.tsx @@ -1,7 +1,7 @@ import { useEffect } from 'react' import type { ExecutionHostId } from '../../../../shared/execution-host' import { useAppStore } from '@/store' -import { installWindowVisibilityInterval } from '@/lib/window-visibility-interval' +import { installWindowVisibilityInterval } from '@/lib/window-activity/window-visibility-interval' import { findWorktreeReviewOwner } from '@/store/github/refresh/worktree-refresh' /** Re-reads the worktree's current linked issue in this window's store; fresh-gated by fetchIssue. */ From 985176b11bbb92175043a8d302581412d703e81c Mon Sep 17 00:00:00 2001 From: Brennan Benson <79079362+brennanb2025@users.noreply.github.com> Date: Tue, 10 Nov 2026 17:08:11 -0800 Subject: [PATCH 15/21] fix(remote-runtime): reopen client-event streams after connection replacement (#40881) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit * fix(remote-runtime): end remote streams when the host connection is replaced A remote stream remembered its host connection's transport generation and, once the generation moved on, main silently dropped every frame without closing the stream. The chat reader still believed it was connected and never reopened, so messages stayed "Sending…" until the tab was reopened. Advancing the generation (including eviction) now ends every pending and retained subscription fenced on the old one; readers resume from their applied cursor on the new connection. Tracked in STA-14427. * fix(remote-runtime): resubscribe client events when their stream ends Main now ends streams whose host connection was replaced, but the renderer's client-event owner ignored the close and kept a dead subscription while its key stayed the same, so repository, worktree and SSH updates stopped. It now drops that subscription on close (also when the close lands mid-setup) and resubscribes through its existing retry. Tracked in STA-14427. * fix(remote-runtime): back off client-event resubscribes until a stream stays up A host that accepted the client-event subscription and then dropped it made the owner resubscribe about once a second forever, because each install reset the failure count. Every close now counts toward the existing capped backoff, and the count resets only once a stream has stayed up for 30 seconds. Tracked in STA-14427. * fix(remote-runtime): retain the landed transport retirement implementation * fix(remote-runtime): isolate client-event retries by host --- .../ipc-events/runtime-client-ipc-bridge.ts | 6 +- ...time-environment-subscription-selection.ts | 4 +- .../hooks/runtime-client-events-sync.test.ts | 169 ++++++++- .../src/hooks/runtime-client-events-sync.ts | 327 +++++++----------- ...ents-runtime-environment-selectors.test.ts | 74 ++++ .../subscriptions/runtime-client-events.ts | 5 +- 6 files changed, 383 insertions(+), 202 deletions(-) diff --git a/src/renderer/src/hooks/ipc-events/runtime-client-ipc-bridge.ts b/src/renderer/src/hooks/ipc-events/runtime-client-ipc-bridge.ts index ea7bdb181c0..3531d02b757 100644 --- a/src/renderer/src/hooks/ipc-events/runtime-client-ipc-bridge.ts +++ b/src/renderer/src/hooks/ipc-events/runtime-client-ipc-bridge.ts @@ -236,7 +236,7 @@ export function registerRuntimeClientIpcBridge( const runtimeClientEventsSync = createRuntimeClientEventsSync({ getDesiredEnvironmentIds: () => getRuntimeClientEventEnvironmentIds(useAppStore.getState()), getSubscriptionKey: (environmentId) => buildRuntimeClientEventEnvironmentKey([environmentId]), - subscribe: (environmentId, onEvent, onError, isCurrent) => { + subscribe: (environmentId, onEvent, onError, isCurrent, onClose) => { const sshGeneration = getEnvironmentSshStateGeneration(environmentId) const runtimeGeneration = getRuntimeEnvironmentConnectionGeneration(environmentId) const runtimeRevision = getRuntimeEnvironmentRevision(environmentId) @@ -271,10 +271,10 @@ export function registerRuntimeClientIpcBridge( useAppStore.getState().markEnvironmentSshStateStale(environmentId), hydrateEnvironmentSshState: () => hydrateRuntimeEnvironmentSshState(environmentId, { force: true }), - sync: runtimeClientEventsSync.sync + sync: () => runtimeClientEventsSync.sync([environmentId]) }) }, - { hostsBrowserPages: hostsBrowserPagesFor(environmentId) } + { hostsBrowserPages: hostsBrowserPagesFor(environmentId), onClose } ) return subscription }, diff --git a/src/renderer/src/hooks/ipc-events/runtime-environment-subscription-selection.ts b/src/renderer/src/hooks/ipc-events/runtime-environment-subscription-selection.ts index e0737cbd63d..6d435008018 100644 --- a/src/renderer/src/hooks/ipc-events/runtime-environment-subscription-selection.ts +++ b/src/renderer/src/hooks/ipc-events/runtime-environment-subscription-selection.ts @@ -111,7 +111,7 @@ type RuntimeEnvironmentStoreSyncSubscriberDeps = { getReachableEnvironmentIds: (state: RuntimeEnvironmentStoreSyncState) => string[] requestProjectRefresh: (environmentId: string) => void markEnvironmentSshStateStale: (environmentId: string) => void - sync: () => void + sync: (recoveryEnvironmentIds?: readonly string[]) => void } export type RuntimeEnvironmentStoreSyncSubscriber = ( @@ -170,7 +170,7 @@ export function createRuntimeEnvironmentStoreSyncSubscriber( } desiredEnvironmentKey = nextDesiredEnvironmentKey reachableEnvironmentKey = nextReachableEnvironmentKey - deps.sync() + deps.sync(refreshEnvironmentIds) } finally { handlingStoreWrite = false } diff --git a/src/renderer/src/hooks/runtime-client-events-sync.test.ts b/src/renderer/src/hooks/runtime-client-events-sync.test.ts index 529ef3340ee..01f11fe17ad 100644 --- a/src/renderer/src/hooks/runtime-client-events-sync.test.ts +++ b/src/renderer/src/hooks/runtime-client-events-sync.test.ts @@ -211,6 +211,73 @@ describe('createRuntimeClientEventsSync', () => { } }) + it('resubscribes on its own when the stream closes, backing off until a stream stays up', async () => { + vi.useFakeTimers() + try { + const closes: (() => void)[] = [] + const unsubscribes: ReturnType[] = [] + const subscribe = vi.fn( + ( + _environmentId: string, + _onEvent: unknown, + _onError: unknown, + _isCurrent: unknown, + onClose?: () => void + ): Promise => { + const unsubscribe = vi.fn() + unsubscribes.push(unsubscribe) + closes.push(() => onClose?.()) + // The first stream closes before its setup settles. + if (closes.length === 1) { + onClose?.() + } + return Promise.resolve({ unsubscribe }) + } + ) + const sync = createRuntimeClientEventsSync({ + getDesiredEnvironmentIds: () => ['A'], + getSubscriptionKey: () => 'A:1', + subscribe, + onEvent: vi.fn(), + retryDelayMs: 10, + retryMaxDelayMs: 40, + random: () => 1 + }) + + sync.sync() + await vi.advanceTimersByTimeAsync(0) + expect(unsubscribes[0]).toHaveBeenCalledTimes(1) + await vi.advanceTimersByTimeAsync(10) + expect(subscribe).toHaveBeenCalledTimes(2) + + // Streams that close right after setup back off 20, 40, then stay capped at 40. + for (const [attempt, delay] of [ + [2, 20], + [3, 40], + [4, 40] + ]) { + closes[attempt - 1]() + await vi.advanceTimersByTimeAsync(delay - 1) + expect(subscribe).toHaveBeenCalledTimes(attempt) + await vi.advanceTimersByTimeAsync(1) + expect(subscribe).toHaveBeenCalledTimes(attempt + 1) + } + + // A late close from a replaced stream must not drop the live one. + closes[1]() + await vi.advanceTimersByTimeAsync(40) + expect(subscribe).toHaveBeenCalledTimes(5) + + // A stream that stayed up 30s ends the streak: the next retry uses the base delay. + await vi.advanceTimersByTimeAsync(30_000) + closes[4]() + await vi.advanceTimersByTimeAsync(10) + expect(subscribe).toHaveBeenCalledTimes(6) + } finally { + vi.useRealTimers() + } + }) + it('backs off exponentially with a cap while an environment keeps failing', async () => { vi.useFakeTimers() try { @@ -282,7 +349,7 @@ describe('createRuntimeClientEventsSync', () => { } }) - it('an external sync retries a waiting environment immediately (recovery path)', async () => { + it('an explicit recovery wake retries a waiting environment immediately', async () => { vi.useFakeTimers() try { const subscribe = vi.fn((): Promise => @@ -302,7 +369,7 @@ describe('createRuntimeClientEventsSync', () => { expect(subscribe).toHaveBeenCalledTimes(1) // Reachability changes retry immediately instead of waiting on backoff. - sync.sync() + sync.sync(['A']) await vi.advanceTimersByTimeAsync(0) expect(subscribe).toHaveBeenCalledTimes(2) } finally { @@ -310,6 +377,104 @@ describe('createRuntimeClientEventsSync', () => { } }) + it.each([10, 100])('keeps retries local to due hosts in a fleet of %i', async (hostCount) => { + vi.useFakeTimers() + try { + const desired = Array.from({ length: hostCount }, (_, i) => `host-${i}`) + const getDesiredEnvironmentIds = vi.fn(() => desired) + const getSubscriptionKey = vi.fn((id: string) => id) + const closes = new Map void>() + const subscribe = vi.fn( + (id: string, _event: unknown, _error: unknown, _current: unknown, close: () => void) => { + closes.set(id, close) + return Promise.resolve({ unsubscribe: vi.fn() }) + } + ) + const sync = createRuntimeClientEventsSync({ + getDesiredEnvironmentIds, + getSubscriptionKey, + subscribe, + onEvent: vi.fn(), + random: () => 1 + }) + sync.sync() + await vi.advanceTimersByTimeAsync(0) + getDesiredEnvironmentIds.mockClear() + getSubscriptionKey.mockClear() + subscribe.mockClear() + + closes.get('host-0')!() + await vi.advanceTimersByTimeAsync(1_000) + expect(subscribe.mock.calls.map(([id]) => id)).toEqual(['host-0']) + expect(getDesiredEnvironmentIds).not.toHaveBeenCalled() + expect(getSubscriptionKey).not.toHaveBeenCalled() + + subscribe.mockClear() + for (const close of closes.values()) { + close() + } + // Every host is due once within 2s; only host-0 has a prior failure. + await vi.advanceTimersByTimeAsync(2_000) + expect(subscribe).toHaveBeenCalledTimes(hostCount) + expect(new Set(subscribe.mock.calls.map(([id]) => id)).size).toBe(hostCount) + expect(getDesiredEnvironmentIds).not.toHaveBeenCalled() + expect(getSubscriptionKey).not.toHaveBeenCalled() + sync.stop() + } finally { + vi.useRealTimers() + } + }) + + it('preserves sibling deadlines across retries, reconciliation, and scoped recovery', async () => { + vi.useFakeTimers() + try { + const closes = new Map void>() + const attempts: { id: string; at: number }[] = [] + const sync = createRuntimeClientEventsSync({ + getDesiredEnvironmentIds: () => ['A', 'B'], + subscribe: (id, _event, _error, _current, close) => { + attempts.push({ id, at: Date.now() }) + closes.set(id, close) + return Promise.resolve({ unsubscribe: vi.fn() }) + }, + onEvent: vi.fn(), + random: () => 1 + }) + vi.setSystemTime(0) + sync.sync() + await vi.advanceTimersByTimeAsync(0) + for (const delay of [1_000, 2_000, 4_000, 8_000, 16_000]) { + closes.get('A')!() + await vi.advanceTimersByTimeAsync(delay) + } + // A is due at 61s. B's first close at 31s must not reset A's deadline. + closes.get('A')!() + closes.get('B')!() + sync.sync() + await vi.advanceTimersByTimeAsync(1_000) + expect(attempts.filter(({ id }) => id === 'A').map(({ at }) => at)).toEqual([ + 0, 1_000, 3_000, 7_000, 15_000, 31_000 + ]) + expect(attempts.filter(({ id }) => id === 'B').map(({ at }) => at)).toEqual([0, 32_000]) + + closes.get('B')!() + sync.sync(['B']) + await vi.advanceTimersByTimeAsync(0) + expect(attempts.filter(({ id }) => id === 'B').map(({ at }) => at)).toEqual([ + 0, 32_000, 32_000 + ]) + await vi.advanceTimersByTimeAsync(28_999) + expect(attempts.filter(({ id }) => id === 'A')).toHaveLength(6) + await vi.advanceTimersByTimeAsync(1) + expect(attempts.filter(({ id }) => id === 'A').map(({ at }) => at)).toEqual([ + 0, 1_000, 3_000, 7_000, 15_000, 31_000, 61_000 + ]) + sync.stop() + } finally { + vi.useRealTimers() + } + }) + it('does not carry a stale failure count when a rejection lands after the env left the set', async () => { vi.useFakeTimers() try { diff --git a/src/renderer/src/hooks/runtime-client-events-sync.ts b/src/renderer/src/hooks/runtime-client-events-sync.ts index 19d60767dfe..9709f6160f9 100644 --- a/src/renderer/src/hooks/runtime-client-events-sync.ts +++ b/src/renderer/src/hooks/runtime-client-events-sync.ts @@ -5,8 +5,8 @@ export type RuntimeClientEventSubscriptionHandle = { } export type RuntimeClientEventsSyncDeps = { - /** Current set of runtime environment ids that should have a live client-event - * subscription. Re-read on every sync and at subscribe-resolution time. */ + /** Current desired hosts. Membership and key changes must synchronously call sync; + * retries and settlements use the ownership that reconciliation established. */ getDesiredEnvironmentIds: () => string[] /** Changes when an environment's transport generation requires a fresh subscription. */ getSubscriptionKey?: (environmentId: string) => string @@ -15,7 +15,9 @@ export type RuntimeClientEventsSyncDeps = { onEvent: (event: RuntimeClientEvent) => void, onError: (error: unknown) => void, /** Use for subscription-side recovery that runs before event delivery. */ - isCurrent: () => boolean + isCurrent: () => boolean, + /** The stream ended (e.g. its host connection was replaced); the owner resubscribes. */ + onClose: () => void ) => Promise onEvent: (environmentId: string, event: RuntimeClientEvent) => void /** Base retry delay; doubles per consecutive failure up to retryMaxDelayMs. */ @@ -26,221 +28,160 @@ export type RuntimeClientEventsSyncDeps = { } export type RuntimeClientEventsSync = { - /** Reconciles live subscriptions to the desired environment set. */ - sync: () => void - /** Tears down all subscriptions and bumps the generation so in-flight - * subscribes resolve into a no-op. */ + /** Reconciles desired hosts; recovery ids explicitly wake those hosts before their retry deadline. */ + sync: (recoveryEnvironmentIds?: readonly string[]) => void + /** Retires all ownership so in-flight subscriptions are released on settlement. */ stop: () => void } -/** Reconcile one subscription per desired environment; remove/re-add can race two in-flight subscribes without a generation change. Keep the first winner and unsubscribe duplicates rather than overwriting and leaking its handle. */ +const STABLE_STREAM_MS = 30_000 + +type SubscriptionAttempt = { + closed: boolean + unsubscribe?: () => void + installedAt?: number +} + +type EnvironmentSubscription = { + key: string + attempt?: SubscriptionAttempt + failures: number + retryTimer?: ReturnType +} + export function createRuntimeClientEventsSync( deps: RuntimeClientEventsSyncDeps ): RuntimeClientEventsSync { - type SubscriptionToken = { key: string; generation: number } - const subscriptions = new Map< - string, - { key: string; token: SubscriptionToken; unsubscribe: () => void } - >() - const pending = new Map() - const retryTimers = new Map>() - const consecutiveFailures = new Map() + const environments = new Map() const retryDelayMs = deps.retryDelayMs ?? 1_000 const retryMaxDelayMs = deps.retryMaxDelayMs ?? 30_000 const random = deps.random ?? Math.random const getSubscriptionKey = deps.getSubscriptionKey ?? ((environmentId: string) => environmentId) - let generation = 0 let syncInvocation = 0 - const clearRetryTimer = (environmentId: string): void => { - const retryTimer = retryTimers.get(environmentId) - if (!retryTimer) { - return + const clearRetry = (entry: EnvironmentSubscription): void => { + if (entry.retryTimer !== undefined) { + clearTimeout(entry.retryTimer) + entry.retryTimer = undefined } - clearTimeout(retryTimer) - retryTimers.delete(environmentId) } - const nextRetryDelayMs = (environmentId: string, subscriptionKey: string): number => { - // Exponential capped jitter paces dead hosts without synchronized retries; external reachability/desired transitions still retry immediately. - const failure = consecutiveFailures.get(environmentId) - const failures = failure?.key === subscriptionKey ? failure.count : 0 - const capped = Math.min(retryDelayMs * 2 ** Math.max(0, failures - 1), retryMaxDelayMs) - return capped * (0.5 + random() * 0.5) + const retire = (entry: EnvironmentSubscription): void => { + clearRetry(entry) + const attempt = entry.attempt + entry.attempt = undefined + attempt?.unsubscribe?.() } - const scheduleRetry = ( - environmentId: string, - subscriptionKey: string, - subscribeGeneration: number - ): void => { - if (retryTimers.has(environmentId)) { + const scheduleRetry = (environmentId: string, entry: EnvironmentSubscription): void => { + if (environments.get(environmentId) !== entry || entry.retryTimer !== undefined) { return } - // Store writes no longer drive retries; keep a bounded timer while failed environments remain desired. - const retryTimer = setTimeout( + const capped = Math.min(retryDelayMs * 2 ** Math.max(0, entry.failures - 1), retryMaxDelayMs) + entry.retryTimer = setTimeout( () => { - retryTimers.delete(environmentId) - if ( - subscribeGeneration !== generation || - !deps.getDesiredEnvironmentIds().includes(environmentId) || - getSubscriptionKey(environmentId) !== subscriptionKey - ) { + entry.retryTimer = undefined + if (environments.get(environmentId) === entry) { + startAttempt(environmentId, entry) + } + }, + capped * (0.5 + random() * 0.5) + ) + } + + const startAttempt = (environmentId: string, entry: EnvironmentSubscription): void => { + if (environments.get(environmentId) !== entry || entry.attempt) { + return + } + clearRetry(entry) + const attempt: SubscriptionAttempt = { closed: false } + entry.attempt = attempt + const ownsAttempt = (): boolean => + environments.get(environmentId) === entry && entry.attempt === attempt + const isCurrent = (): boolean => ownsAttempt() && !attempt.closed + void deps + .subscribe( + environmentId, + (event) => { + if (isCurrent()) { + deps.onEvent(environmentId, event) + } + }, + (error) => console.warn('[runtime-client-events] subscription error:', error), + isCurrent, + () => { + attempt.closed = true + if (!ownsAttempt() || attempt.installedAt === undefined) { + return + } + entry.attempt = undefined + if (Date.now() - attempt.installedAt >= STABLE_STREAM_MS) { + entry.failures = 0 + } + entry.failures += 1 + scheduleRetry(environmentId, entry) + } + ) + .then((subscription) => { + if (!ownsAttempt()) { + subscription.unsubscribe() return } - sync() - }, - nextRetryDelayMs(environmentId, subscriptionKey) - ) - retryTimers.set(environmentId, retryTimer) + if (attempt.closed) { + entry.attempt = undefined + subscription.unsubscribe() + entry.failures += 1 + scheduleRetry(environmentId, entry) + return + } + attempt.unsubscribe = subscription.unsubscribe + attempt.installedAt = Date.now() + }) + .catch((error) => { + if (ownsAttempt()) { + entry.attempt = undefined + console.warn('[runtime-client-events] failed to subscribe:', error) + entry.failures += 1 + scheduleRetry(environmentId, entry) + } + }) + } + + const sync = (recoveryEnvironmentIds: readonly string[] = []): void => { + const invocation = ++syncInvocation + const desiredIds = new Set(deps.getDesiredEnvironmentIds()) + for (const [environmentId, entry] of environments) { + if (invocation !== syncInvocation) { + return + } + if (desiredIds.has(environmentId) && entry.key === getSubscriptionKey(environmentId)) { + continue + } + environments.delete(environmentId) + retire(entry) + } + const recoveryIds = new Set(recoveryEnvironmentIds) + for (const environmentId of desiredIds) { + if (invocation !== syncInvocation) { + return + } + let entry = environments.get(environmentId) + if (!entry) { + entry = { key: getSubscriptionKey(environmentId), failures: 0 } + environments.set(environmentId, entry) + } + if (entry.retryTimer === undefined || recoveryIds.has(environmentId)) { + startAttempt(environmentId, entry) + } + } } const stop = (): void => { - generation += 1 - const stoppedSubscriptions = [...subscriptions.values()] - subscriptions.clear() - pending.clear() - for (const retryTimer of retryTimers.values()) { - clearTimeout(retryTimer) - } - retryTimers.clear() - consecutiveFailures.clear() - for (const subscription of stoppedSubscriptions) { - subscription.unsubscribe() - } - } - - const sync = (): void => { - const syncGeneration = generation - const currentSyncInvocation = ++syncInvocation - const desiredIds = new Set(deps.getDesiredEnvironmentIds()) - for (const environmentId of retryTimers.keys()) { - if (desiredIds.has(environmentId)) { - continue - } - clearRetryTimer(environmentId) - } - for (const environmentId of consecutiveFailures.keys()) { - if (!desiredIds.has(environmentId)) { - consecutiveFailures.delete(environmentId) - } - } - - for (const [environmentId, subscription] of subscriptions) { - if (syncGeneration !== generation || currentSyncInvocation !== syncInvocation) { - return - } - if (desiredIds.has(environmentId) && subscription.key === getSubscriptionKey(environmentId)) { - continue - } - subscriptions.delete(environmentId) - subscription.unsubscribe() - } - - for (const environmentId of desiredIds) { - if (syncGeneration !== generation || currentSyncInvocation !== syncInvocation) { - return - } - const subscriptionKey = getSubscriptionKey(environmentId) - const pendingSubscription = pending.get(environmentId) - if (pendingSubscription && pendingSubscription.key !== subscriptionKey) { - pending.delete(environmentId) - } - if ( - subscriptions.get(environmentId)?.key === subscriptionKey || - pending.get(environmentId)?.key === subscriptionKey - ) { - continue - } - clearRetryTimer(environmentId) - const subscribeGeneration = generation - const pendingSubscriptionToken = { key: subscriptionKey, generation: subscribeGeneration } - pending.set(environmentId, pendingSubscriptionToken) - const isCurrent = (): boolean => - subscribeGeneration === generation && - (pending.get(environmentId) === pendingSubscriptionToken || - subscriptions.get(environmentId)?.token === pendingSubscriptionToken) - void deps - .subscribe( - environmentId, - (event) => { - // Initial frames arrive before setup settles; only the owning attempt may deliver. - if (isCurrent()) { - deps.onEvent(environmentId, event) - } - }, - (error) => { - console.warn('[runtime-client-events] subscription error:', error) - }, - isCurrent - ) - .then((subscription) => { - const isCurrentPending = pending.get(environmentId) === pendingSubscriptionToken - if (isCurrentPending) { - pending.delete(environmentId) - } - if ( - !isCurrentPending || - subscribeGeneration !== generation || - !deps.getDesiredEnvironmentIds().includes(environmentId) || - getSubscriptionKey(environmentId) !== subscriptionKey - ) { - subscription.unsubscribe() - return - } - // Another subscribe already won; unsubscribe this duplicate instead of overwriting and leaking the winner’s handle. - if (subscriptions.get(environmentId)?.key === subscriptionKey) { - subscription.unsubscribe() - return - } - consecutiveFailures.delete(environmentId) - subscriptions.set(environmentId, { - key: subscriptionKey, - token: pendingSubscriptionToken, - unsubscribe: subscription.unsubscribe - }) - }) - .catch((error) => { - const isCurrentPending = pending.get(environmentId) === pendingSubscriptionToken - if (isCurrentPending) { - pending.delete(environmentId) - } - if ( - isCurrentPending && - subscribeGeneration === generation && - getSubscriptionKey(environmentId) === subscriptionKey - ) { - console.warn('[runtime-client-events] failed to subscribe:', error) - // Count failures only while retrying this desired environment; late rejection after removal must not penalize re-entry. - if (deps.getDesiredEnvironmentIds().includes(environmentId)) { - const failure = consecutiveFailures.get(environmentId) - consecutiveFailures.set(environmentId, { - key: subscriptionKey, - count: failure?.key === subscriptionKey ? failure.count + 1 : 1 - }) - scheduleRetry(environmentId, subscriptionKey, subscribeGeneration) - } else { - consecutiveFailures.delete(environmentId) - } - } - }) - } - - if (syncGeneration !== generation || currentSyncInvocation !== syncInvocation) { - return - } - for (const [environmentId, pendingSubscription] of pending) { - if ( - desiredIds.has(environmentId) && - pendingSubscription.key === getSubscriptionKey(environmentId) - ) { - continue - } - pending.delete(environmentId) - } - - if (desiredIds.size === 0 && subscriptions.size === 0) { - generation += 1 + syncInvocation += 1 + const stopped = [...environments.values()] + environments.clear() + for (const entry of stopped) { + retire(entry) } } diff --git a/src/renderer/src/hooks/useIpcEvents-runtime-environment-selectors.test.ts b/src/renderer/src/hooks/useIpcEvents-runtime-environment-selectors.test.ts index a789a42c195..461640eda60 100644 --- a/src/renderer/src/hooks/useIpcEvents-runtime-environment-selectors.test.ts +++ b/src/renderer/src/hooks/useIpcEvents-runtime-environment-selectors.test.ts @@ -10,6 +10,8 @@ import type { RuntimeEnvironmentStoreSyncSubscriber } from './ipc-events/runtime-environment-subscription-selection' +import { createRuntimeClientEventsSync } from './runtime-client-events-sync' + type RuntimeEnvironmentStoreSubscriber = RuntimeEnvironmentStoreSyncSubscriber type RuntimeEnvironmentStoreState = RuntimeEnvironmentStoreSyncState @@ -32,6 +34,78 @@ function makeRuntimeEnvironmentStoreState(args: { } describe('createRuntimeEnvironmentStoreSyncSubscriber', () => { + it('retires removed registry members despite retained status and follows active host changes', async () => { + const environments = [ + { id: 'A', createdAt: 1 }, + { id: 'B', createdAt: 2 } + ] + const statuses = new Map([ + ['A', { status: { runtimeId: 'peer-a' }, checkedAt: 1 }], + ['B', { status: { runtimeId: 'peer-b' }, checkedAt: 1 }] + ]) + let state = makeRuntimeEnvironmentStoreState({ environments, statuses }) + const subscriptions: string[] = [] + const retirements: string[] = [] + const wakes: (readonly string[] | undefined)[] = [] + const manager = createRuntimeClientEventsSync({ + getDesiredEnvironmentIds: () => getRuntimeClientEventEnvironmentIds(state), + subscribe: (id) => { + subscriptions.push(id) + return Promise.resolve({ + unsubscribe: () => { + retirements.push(id) + } + }) + }, + onEvent: vi.fn() + }) + const subscriber = createRuntimeEnvironmentStoreSyncSubscriber({ + initialDesiredEnvironmentIds: getRuntimeClientEventEnvironmentIds(state), + initialReachableEnvironmentIds: getReachableRuntimeEnvironmentIds(state), + getDesiredEnvironmentIds: getRuntimeClientEventEnvironmentIds, + getReachableEnvironmentIds: getReachableRuntimeEnvironmentIds, + buildEnvironmentKey: (ids) => [...ids].sort().join(':'), + requestProjectRefresh: vi.fn(), + markEnvironmentSshStateStale: vi.fn(), + sync: (ids) => { + wakes.push(ids) + manager.sync(ids) + } + }) + const publish = (next: RuntimeEnvironmentStoreState): void => { + const previous = state + state = next + subscriber(state, previous) + } + manager.sync() + await Promise.resolve() + publish(makeRuntimeEnvironmentStoreState({ environments: [environments[1]], statuses })) + expect(retirements).toEqual(['A']) + expect(wakes).toEqual([[]]) + publish( + makeRuntimeEnvironmentStoreState({ + environments: [environments[1]], + statuses, + activeEnvironmentId: 'C' + }) + ) + await Promise.resolve() + expect(subscriptions).toEqual(['A', 'B', 'C']) + expect(wakes).toEqual([[], ['C']]) + publish( + makeRuntimeEnvironmentStoreState({ + environments: [environments[1]], + statuses, + activeEnvironmentId: 'D' + }) + ) + await Promise.resolve() + expect(retirements).toEqual(['A', 'C']) + expect(subscriptions).toEqual(['A', 'B', 'C', 'D']) + expect(wakes).toEqual([[], ['C'], ['D']]) + manager.stop() + }) + it('syncs each connect, disconnect, re-pair, runtime generation, and nested SSH generation once', () => { let environments = [ { id: 'runtime-a', createdAt: 1, pairingRevision: 1 }, diff --git a/src/renderer/src/runtime/rpc/subscriptions/runtime-client-events.ts b/src/renderer/src/runtime/rpc/subscriptions/runtime-client-events.ts index 0c407253fc8..e461cc6284c 100644 --- a/src/renderer/src/runtime/rpc/subscriptions/runtime-client-events.ts +++ b/src/renderer/src/runtime/rpc/subscriptions/runtime-client-events.ts @@ -21,7 +21,7 @@ export async function subscribeRuntimeClientEvents( // Control-outage events are lost, not queued. First replay-tagged response is the only signal to resync event-derived mirrors such as per-environment SSH state. onReplayedAfterReconnect?: () => void, // Managed SSH hosts reach this desktop's browser through the desktop browser bridge instead. - options: { hostsBrowserPages: boolean } = { hostsBrowserPages: false } + options: { hostsBrowserPages: boolean; onClose?: () => void } = { hostsBrowserPages: false } ): Promise { const handle = await subscribeRuntimeEnvironment( { @@ -42,7 +42,8 @@ export async function subscribeRuntimeClientEvents( onReplayedAfterReconnect ) }, - onError + onError, + onClose: options.onClose } ) return { unsubscribe: handle.unsubscribe } From 8aa3f9c36e30d045252646768e84de9d81b6f12c Mon Sep 17 00:00:00 2001 From: Neil <4138956+nwparker@users.noreply.github.com> Date: Tue, 10 Nov 2026 17:17:59 -0800 Subject: [PATCH 16/21] refactor(renderer): group browser owners and typing diagnostics (#40936) --- config/flaky-test-pattern-baseline.txt | 1 - config/renderer-layout-writer-baseline.txt | 2 +- .../scripts/ci/sharding/ci-shard-timings.json | 15 +- config/workspace-record-read-baseline.txt | 2 +- .../BrowserPaneOverlayLayer.tsx | 2 +- ...er-workspace-pane.retention-props.test.tsx | 5 +- .../browser-workspace-pane.tsx | 4 +- .../browser-client-page-paint-retention.ts | 2 +- .../browser-guest-eviction-veto.test.ts | 76 --- .../browser-guest-paint-retention.ts | 8 +- .../use-checks-panel-git-status-effects.tsx | 2 +- .../sync/git-status-refresh-scheduler.ts | 2 +- .../sync/stale-conflict-operation-poll.ts | 2 +- .../source-control/sync/use-branch-compare.ts | 2 +- .../sidebar-desktop-facts.test-fixture.ts} | 6 +- .../sidebar-list-parity.test-fixture.ts | 6 +- ...idebar-paired-server-rows.test-fixture.ts} | 24 +- ...idebar-workspace-activity.test-fixture.ts} | 8 +- .../sidebar-workspace-activity.test.ts} | 4 +- .../tab-bar/ClientHostedBrowserTabRows.tsx | 2 +- .../src/components/tab-bar/TabBar.tsx | 2 +- .../components/tab-bar/TabBarItemRow.test.tsx | 271 --------- .../tab-bar/use-tab-bar-item-actions.ts | 2 +- .../TabGroupPanel.os-file-drop.test.tsx | 118 ---- .../TabGroupPanel.pane-actions.test.tsx | 86 --- .../components/tab-group/TabGroupPanel.tsx | 2 +- .../use-terminal-workspace-projection.ts | 2 +- .../typing-latency/census-probe.ts | 7 +- .../typing-latency/diagnostic-report.ts | 12 +- .../typing-latency/diagnostic-summary.ts | 0 .../typing-latency/diagnostic.ts | 10 +- .../typing-latency/echo-candidate-timeout.ts | 0 .../echo-instrumentation.test.ts | 204 +++++++ .../typing-latency/echo-instrumentation.ts | 38 +- .../typing-latency/echo-observation.ts | 0 .../typing-latency/input-events.ts | 0 .../typing-latency/input-source.ts | 9 +- .../typing-latency/pane-target.ts | 0 .../renderer-long-tasks.test.ts | 0 .../typing-latency/renderer-long-tasks.ts | 0 .../typing-latency/sample-window.ts | 0 .../browser-remote-viewer-ipc-bridge.ts | 2 +- .../ipc-events/mobile-driver-ipc-bridge.ts | 4 +- ...ts-agent-status-snapshot-hydration.test.ts | 354 ------------ ...cEvents-client-hosted-browser-rows.test.ts | 72 --- .../structured-agent-session-launch-cancel.ts | 2 +- ...tured-agent-session-launch-cancellation.ts | 4 +- ...tructured-agent-session-launch-registry.ts | 2 +- ...-agent-session-launch-cancellation.test.ts | 396 ------------- .../coalesced-poll-runner.test.ts | 319 ----------- .../refresh}/coalesced-poll-runner.ts | 0 .../echo-instrumentation.test.ts | 534 ------------------ .../lib/typing-latency/sample-window.test.ts | 90 --- ...space-tab-palette-activation.store.test.ts | 2 +- .../tabs/workspace-tab-palette-activation.ts | 2 +- src/renderer/src/main.tsx | 2 +- .../client-hosted-browser-row-state.ts | 2 +- .../ownership}/browser-mobile-driver-state.ts | 2 +- .../ownership}/browser-remote-viewer-state.ts | 9 - .../ownership}/browser-remote-viewer-store.ts | 0 .../inventory-refresh.ts | 2 +- .../snapshot-apply.ts | 2 +- .../hosted-review-request-state.ts | 2 +- tests/e2e/typing-scale-census.ts | 6 +- 64 files changed, 297 insertions(+), 2451 deletions(-) delete mode 100644 src/renderer/src/components/browser-pane/host-guest/browser-guest-eviction-veto.test.ts rename src/renderer/src/{lib/sidebar-shadow/desktop-facts-shadow.ts => components/sidebar/worktree-list/parity/sidebar-desktop-facts.test-fixture.ts} (89%) rename src/renderer/src/{lib/sidebar-shadow/paired-server-rows-shadow.ts => components/sidebar/worktree-list/parity/sidebar-paired-server-rows.test-fixture.ts} (76%) rename src/renderer/src/{lib/sidebar-shadow/workspace-activity-shadow.ts => components/sidebar/worktree-list/parity/sidebar-workspace-activity.test-fixture.ts} (92%) rename src/renderer/src/{lib/sidebar-shadow/workspace-activity-shadow.test.ts => components/sidebar/worktree-list/parity/sidebar-workspace-activity.test.ts} (88%) delete mode 100644 src/renderer/src/components/tab-bar/TabBarItemRow.test.tsx delete mode 100644 src/renderer/src/components/tab-group/TabGroupPanel.os-file-drop.test.tsx delete mode 100644 src/renderer/src/components/tab-group/TabGroupPanel.pane-actions.test.tsx rename src/renderer/src/{lib => diagnostics}/typing-latency/census-probe.ts (94%) rename src/renderer/src/{lib => diagnostics}/typing-latency/diagnostic-report.ts (94%) rename src/renderer/src/{lib => diagnostics}/typing-latency/diagnostic-summary.ts (100%) rename src/renderer/src/{lib => diagnostics}/typing-latency/diagnostic.ts (97%) rename src/renderer/src/{lib => diagnostics}/typing-latency/echo-candidate-timeout.ts (100%) create mode 100644 src/renderer/src/diagnostics/typing-latency/echo-instrumentation.test.ts rename src/renderer/src/{lib => diagnostics}/typing-latency/echo-instrumentation.ts (91%) rename src/renderer/src/{lib => diagnostics}/typing-latency/echo-observation.ts (100%) rename src/renderer/src/{lib => diagnostics}/typing-latency/input-events.ts (100%) rename src/renderer/src/{lib => diagnostics}/typing-latency/input-source.ts (89%) rename src/renderer/src/{lib => diagnostics}/typing-latency/pane-target.ts (100%) rename src/renderer/src/{lib => diagnostics}/typing-latency/renderer-long-tasks.test.ts (100%) rename src/renderer/src/{lib => diagnostics}/typing-latency/renderer-long-tasks.ts (100%) rename src/renderer/src/{lib => diagnostics}/typing-latency/sample-window.ts (100%) delete mode 100644 src/renderer/src/hooks/useIpcEvents-agent-status-snapshot-hydration.test.ts delete mode 100644 src/renderer/src/hooks/useIpcEvents-client-hosted-browser-rows.test.ts rename src/renderer/src/lib/{launch-cancel-tombstones-compat => agent-session/launch}/structured-agent-session-launch-cancellation.ts (97%) delete mode 100644 src/renderer/src/lib/launch-cancel-tombstones-compat/structured-agent-session-launch-cancellation.test.ts delete mode 100644 src/renderer/src/lib/repository-refresh/coalesced-poll-runner.test.ts rename src/renderer/src/lib/{repository-refresh => source-control/refresh}/coalesced-poll-runner.ts (100%) delete mode 100644 src/renderer/src/lib/typing-latency/echo-instrumentation.test.ts delete mode 100644 src/renderer/src/lib/typing-latency/sample-window.test.ts rename src/renderer/src/{lib/pane-manager => runtime/browser/client-hosting}/client-hosted-browser-row-state.ts (98%) rename src/renderer/src/{lib/pane-manager => runtime/browser/ownership}/browser-mobile-driver-state.ts (97%) rename src/renderer/src/{lib/pane-manager => runtime/browser/ownership}/browser-remote-viewer-state.ts (81%) rename src/renderer/src/{lib/pane-manager => runtime/browser/ownership}/browser-remote-viewer-store.ts (100%) diff --git a/config/flaky-test-pattern-baseline.txt b/config/flaky-test-pattern-baseline.txt index 362d8073cc6..6820888e95d 100644 --- a/config/flaky-test-pattern-baseline.txt +++ b/config/flaky-test-pattern-baseline.txt @@ -29,7 +29,6 @@ 1 test-no-in-test-import-actual:src/renderer/src/components/terminal-pane/pty-connection/notifications/pty-connection-hook-idle-arbitration.test.ts 3 test-no-in-test-import-actual:src/renderer/src/components/terminal-pane/snapshots/remote-snapshot-alt-screen-host-chain.test.ts 1 test-no-in-test-import-actual:src/renderer/src/hooks/useIpcEvents-agent-status-pane-teardown.test.ts -1 test-no-in-test-import-actual:src/renderer/src/hooks/useIpcEvents-agent-status-snapshot-hydration.test.ts 1 test-no-polling-loop:src/main/browser/cookies/browser-cookie-import-concurrency.test.ts 1 test-no-polling-loop:src/main/claude/claude-stop-ack-race.test.ts 1 test-no-polling-loop:src/main/claude/claude-stream-json-connection.test.ts diff --git a/config/renderer-layout-writer-baseline.txt b/config/renderer-layout-writer-baseline.txt index 750730bd624..26d3f04ab48 100644 --- a/config/renderer-layout-writer-baseline.txt +++ b/config/renderer-layout-writer-baseline.txt @@ -18,7 +18,7 @@ 3 src/renderer/src/components/worktree-jump/worktree-jump-palette-status-inputs.ts 1 src/renderer/src/lib/agent-session/hibernation/agent-hibernation-coordinator.ts 4 src/renderer/src/components/floating-terminal/floating-workspace-terminal-actions.ts -1 src/renderer/src/lib/pane-manager/client-hosted-browser-row-state.ts +1 src/renderer/src/runtime/browser/client-hosting/client-hosted-browser-row-state.ts 4 src/renderer/src/lib/workspace/persistence/session-write-subscriber.ts 1 src/renderer/src/lib/workspace/persistence/workspace-session-browser-tabs.ts 7 src/renderer/src/lib/workspace/persistence/workspace-session-host-hydration.ts diff --git a/config/scripts/ci/sharding/ci-shard-timings.json b/config/scripts/ci/sharding/ci-shard-timings.json index b583ee3ab62..797e552a50f 100644 --- a/config/scripts/ci/sharding/ci-shard-timings.json +++ b/config/scripts/ci/sharding/ci-shard-timings.json @@ -4744,7 +4744,6 @@ "src/renderer/src/components/browser-pane/history/use-browser-history-shortcut.test.tsx": 1990, "src/renderer/src/components/browser-pane/host-guest/agent-input-focus-guard.test.ts": 228, "src/renderer/src/components/browser-pane/host-guest/browser-automation-visibility.test.ts": 45, - "src/renderer/src/components/browser-pane/host-guest/browser-guest-eviction-veto.test.ts": 1724, "src/renderer/src/components/browser-pane/host-guest/browser-page-evicted-guest-recovery.test.ts": 1839, "src/renderer/src/components/browser-pane/host-guest/browser-page-guest-recovery.test.ts": 515, "src/renderer/src/components/browser-pane/host-guest/browser-page-registration-ownership.test.ts": 2197, @@ -5366,7 +5365,6 @@ "src/renderer/src/components/right-sidebar/checks-panel/gitlab-review-client.test.ts": 321, "src/renderer/src/components/right-sidebar/checks-panel/comments/review-comment-markdown.test.tsx": 2171, "src/renderer/src/components/right-sidebar/checks-panel/refresh/use-checks-panel-manual-refresh-recovery.test.tsx": 2909, - "src/renderer/src/lib/repository-refresh/coalesced-poll-runner.test.ts": 61, "src/renderer/src/components/right-sidebar/file-explorer/mutations/file-explorer-delete-undo-precondition.test.tsx": 2023, "src/renderer/src/components/right-sidebar/file-explorer/tree/file-explorer-dir-pages.test.ts": 1767, "src/renderer/src/components/right-sidebar/file-explorer/tree/file-explorer-directory-listing.test.ts": 2072, @@ -5631,7 +5629,6 @@ "src/renderer/src/components/status-bar/use-resource-usage-derived-model.test.tsx": 742, "src/renderer/src/components/tab-bar/CopyTabIdMenuItem.test.tsx": 428, "src/renderer/src/components/tab-bar/TabBarCreateEntry.search.test.tsx": 3224, - "src/renderer/src/components/tab-bar/TabBarItemRow.test.tsx": 2116, "src/renderer/src/components/tab-bar/TabBarQuickCommandItem.test.tsx": 384, "src/renderer/src/components/tab-bar/TabBarQuickCommandsMenu.virtualization.test.tsx": 1644, "src/renderer/src/components/tab-bar/TabStripScrollIndicator.test.tsx": 356, @@ -5656,8 +5653,6 @@ "src/renderer/src/components/tab-bar/use-sortable-tab-context-menu.test.ts": 261, "src/renderer/src/components/tab-bar/use-tab-bar-quick-command-search-input.test.tsx": 536, "src/renderer/src/components/tab-group/RetainedPaneHost.test.tsx": 340, - "src/renderer/src/components/tab-group/TabGroupPanel.os-file-drop.test.tsx": 1711, - "src/renderer/src/components/tab-group/TabGroupPanel.pane-actions.test.tsx": 1419, "src/renderer/src/components/tab-group/tab-drag-pointer-sensor.test.ts": 187, "src/renderer/src/components/tab-group/tab-drag-preview-activation.test.ts": 1785, "src/renderer/src/components/tab-group/tab-drag-retained-guest-passthrough.test.tsx": 2104, @@ -6118,12 +6113,10 @@ "src/renderer/src/hooks/useIpcEvents-agent-status-connection-attribution.test.ts": 5510, "src/renderer/src/hooks/useIpcEvents-agent-status-pane-teardown.test.ts": 5268, "src/renderer/src/hooks/useIpcEvents-agent-status-queue-ordering.test.ts": 3680, - "src/renderer/src/hooks/useIpcEvents-agent-status-snapshot-hydration.test.ts": 7403, "src/renderer/src/hooks/useIpcEvents-agent-status-snapshot-replay.test.ts": 4119, "src/renderer/src/hooks/useIpcEvents-agent-status-ssh-authority.test.ts": 3310, "src/renderer/src/hooks/useIpcEvents-browser-certificate-failure.test.ts": 2981, "src/renderer/src/hooks/useIpcEvents-browser-navigation.test.ts": 1511, - "src/renderer/src/hooks/useIpcEvents-client-hosted-browser-rows.test.ts": 1894, "src/renderer/src/hooks/useIpcEvents-close-routing-session-tabs.test.ts": 5812, "src/renderer/src/hooks/useIpcEvents-lifecycle.test.ts": 2326, "src/renderer/src/hooks/useIpcEvents-new-workspace-shortcut.test.ts": 72, @@ -6254,7 +6247,6 @@ "src/renderer/src/lib/agent-session/launch/new-tab/launch-agent-new-tab-host-route-readiness.test.ts": 489, "src/renderer/src/lib/agent-session/launch/new-tab/launch-agent-picker-contract.test.ts": 3592, "src/renderer/src/lib/agent-session/launch/continuation/launch-agent-session-continuation-toast.test.tsx": 2347, - "src/renderer/src/lib/launch-cancel-tombstones-compat/structured-agent-session-launch-cancellation.test.ts": 2266, "src/renderer/src/lib/workspace/creation/work-item/launch-work-item-direct-accepted.test.tsx": 2754, "src/renderer/src/lib/workspace/creation/work-item/launch-work-item-direct-defaults.test.ts": 1594, "src/renderer/src/lib/workspace/creation/work-item/launch-work-item-direct-setup.test.tsx": 3435, @@ -6359,7 +6351,7 @@ "src/renderer/src/lib/review-owner.test.ts": 238, "src/renderer/src/lib/runtime-host-copy.test.ts": 317, "src/renderer/src/components/sidebar/sidebar-feed/sidebar-pending-layer.test.ts": 338, - "src/renderer/src/lib/sidebar-shadow/workspace-activity-shadow.test.ts": 1728, + "src/renderer/src/components/sidebar/worktree-list/parity/sidebar-workspace-activity.test.ts": 1728, "src/renderer/src/lib/simulator/launch/simulator-launch-coordination.test.ts": 43, "src/renderer/src/lib/simulator/search/simulator-palette-search.test.ts": 515, "src/renderer/src/lib/simulator/tabs/simulator-pane-shutdown-scheduler.test.ts": 1679, @@ -6379,9 +6371,8 @@ "src/renderer/src/lib/text-input/paste/text-control-paste-ownership.test.ts": 217, "src/renderer/src/lib/text-input/paste/text-control-paste.test.ts": 222, "src/renderer/src/lib/text-generation-error-copy.test.ts": 466, - "src/renderer/src/lib/typing-latency/echo-instrumentation.test.ts": 145, - "src/renderer/src/lib/typing-latency/renderer-long-tasks.test.ts": 75, - "src/renderer/src/lib/typing-latency/sample-window.test.ts": 77, + "src/renderer/src/diagnostics/typing-latency/echo-instrumentation.test.ts": 145, + "src/renderer/src/diagnostics/typing-latency/renderer-long-tasks.test.ts": 75, "src/renderer/src/attention/unread/unread-badge-count-selector.test.ts": 1930, "src/renderer/src/lib/agent-session/tab-identity/use-tab-agent-observed-signal-dispatch.test.tsx": 1971, "src/renderer/src/lib/agent-session/tab-identity/use-tab-agent.test.ts": 1961, diff --git a/config/workspace-record-read-baseline.txt b/config/workspace-record-read-baseline.txt index f6555eb8c82..b6e4e4e16f9 100644 --- a/config/workspace-record-read-baseline.txt +++ b/config/workspace-record-read-baseline.txt @@ -36,7 +36,7 @@ 1 src/renderer/src/lib/workspace/ownership/resolved-worktree-execution-host.ts # P5b-2 must clear: loosely typed state param 2 src/renderer/src/lib/workspace/ownership/runtime-session-mirror-owners.ts # P5b-2 must clear: optional state fields default to empty 2 src/renderer/src/lib/workspace/ownership/terminal-worktree-route.ts # P5b-2 must clear: the unattributed-twin check reads both catalogs directly -3 src/renderer/src/lib/typing-latency/diagnostic-summary.ts # P5b-2 must clear: optional state field +3 src/renderer/src/diagnostics/typing-latency/diagnostic-summary.ts # P5b-2 must clear: optional state field 2 src/renderer/src/attention/unread/unread-badge-count.ts # reads a passed-in copy 1 src/renderer/src/lib/workspace/paths/workspace-directory.ts # P5b-2 must clear: optional state field defaults to empty 2 src/renderer/src/lib/workspace/ownership/workspace-terminal-host-authority.ts # P5b-2 must clear: optional state fields diff --git a/src/renderer/src/components/browser-pane/assemble-chrome/BrowserPaneOverlayLayer.tsx b/src/renderer/src/components/browser-pane/assemble-chrome/BrowserPaneOverlayLayer.tsx index 115ff430fc4..a7e89388e8e 100644 --- a/src/renderer/src/components/browser-pane/assemble-chrome/BrowserPaneOverlayLayer.tsx +++ b/src/renderer/src/components/browser-pane/assemble-chrome/BrowserPaneOverlayLayer.tsx @@ -15,7 +15,7 @@ import { isClientHostedBrowserRowSelectionLive, useClientHostedBrowserRowSelection, useClientHostedBrowserRows -} from '@/lib/pane-manager/client-hosted-browser-row-state' +} from '@/runtime/browser/client-hosting/client-hosted-browser-row-state' import { ClientHostedBrowserHostRowPane } from '../client-hosted-browser-host-row-pane' import { useAnyBrowserPageMountAdmission } from '../host-guest/browser-page-mount-admission' diff --git a/src/renderer/src/components/browser-pane/assemble-chrome/browser-workspace-pane.retention-props.test.tsx b/src/renderer/src/components/browser-pane/assemble-chrome/browser-workspace-pane.retention-props.test.tsx index b612b0ad49e..f3e8196de32 100644 --- a/src/renderer/src/components/browser-pane/assemble-chrome/browser-workspace-pane.retention-props.test.tsx +++ b/src/renderer/src/components/browser-pane/assemble-chrome/browser-workspace-pane.retention-props.test.tsx @@ -69,8 +69,9 @@ vi.mock('./browser-page-pane', () => ({ return } })) -import { hydrateBrowserDrivers } from '@/lib/pane-manager/browser-mobile-driver-state' -import { hydrateBrowserRemoteViewerPages } from '@/lib/pane-manager/browser-remote-viewer-state' +import { hydrateBrowserDrivers } from '@/runtime/browser/ownership/browser-mobile-driver-state' +import '@/runtime/browser/ownership/browser-remote-viewer-state' +import { hydrateBrowserRemoteViewerPages } from '@/runtime/browser/ownership/browser-remote-viewer-store' import BrowserPane from './browser-workspace-pane' const WORKSPACE_ID = 'ws-1' diff --git a/src/renderer/src/components/browser-pane/assemble-chrome/browser-workspace-pane.tsx b/src/renderer/src/components/browser-pane/assemble-chrome/browser-workspace-pane.tsx index 1debc6b62af..cdc6ffbea6e 100644 --- a/src/renderer/src/components/browser-pane/assemble-chrome/browser-workspace-pane.tsx +++ b/src/renderer/src/components/browser-pane/assemble-chrome/browser-workspace-pane.tsx @@ -11,8 +11,8 @@ import { IDLE_BROWSER_DRIVER, useBrowserDriverForPage, useBrowserMobileDrivenPageIds -} from '@/lib/pane-manager/browser-mobile-driver-state' -import { useBrowserRemotelyViewedPageIds } from '@/lib/pane-manager/browser-remote-viewer-state' +} from '@/runtime/browser/ownership/browser-mobile-driver-state' +import { useBrowserRemotelyViewedPageIds } from '@/runtime/browser/ownership/browser-remote-viewer-state' import { useContextualTour } from '@/components/contextual-tours/use-contextual-tour' import { getBrowserPageRuntimeEnvironmentId } from '../describe-page/browser-page-url-display' import type { BrowserChromeShortcutScope } from '../describe-page/browser-page-types' diff --git a/src/renderer/src/components/browser-pane/browser-client-page-paint-retention.ts b/src/renderer/src/components/browser-pane/browser-client-page-paint-retention.ts index 0fac4a7fcb3..0f8adde5afb 100644 --- a/src/renderer/src/components/browser-pane/browser-client-page-paint-retention.ts +++ b/src/renderer/src/components/browser-pane/browser-client-page-paint-retention.ts @@ -3,7 +3,7 @@ import { isWebviewDragPassthroughActive } from './host-guest/webview-drag-passth import { isBrowserPageRemotelyViewed, onBrowserRemoteViewerChange -} from '../../lib/pane-manager/browser-remote-viewer-store' +} from '../../runtime/browser/ownership/browser-remote-viewer-store' import type { BrowserClientRetainedRendererPage } from './browser-client-page-retained-state' export function syncBrowserClientPagePaintRetention(page: BrowserClientRetainedRendererPage): void { diff --git a/src/renderer/src/components/browser-pane/host-guest/browser-guest-eviction-veto.test.ts b/src/renderer/src/components/browser-pane/host-guest/browser-guest-eviction-veto.test.ts deleted file mode 100644 index fcd36d8603f..00000000000 --- a/src/renderer/src/components/browser-pane/host-guest/browser-guest-eviction-veto.test.ts +++ /dev/null @@ -1,76 +0,0 @@ -import { beforeEach, describe, expect, it, vi } from 'vitest' -import type { BrowserWorkspace } from '../../../../../shared/browser-workspace-types' -import { browserTabsVetoGuestEviction } from './browser-guest-eviction-veto' -import { selectBrowserGuestEvictionWorktreeIds } from './browser-guest-worktree-retention' -import { hydrateBrowserDrivers } from '../../../lib/pane-manager/browser-mobile-driver-state' -import { hydrateBrowserRemoteViewerPages } from '../../../lib/pane-manager/browser-remote-viewer-state' - -import { installBrowserPageDownloadActivityTracking } from '../navigate/browser-page-download-activity' - -// Eviction destroys guests, permanently ending paired-client streams; paint retention must veto it too. -const WATCHED_PAGE = 'page-watched' -const WATCHED_WORKTREE = 'wt-watched' - -function tabsFor(pageId: string): BrowserWorkspace[] { - return [{ id: 'tab-1', pageIds: [pageId] } as unknown as BrowserWorkspace] -} - -/** Six over-budget hidden worktrees, so only the veto can spare the watched one. */ -function evictionRun(tabs: readonly BrowserWorkspace[]): string[] { - return selectBrowserGuestEvictionWorktreeIds({ - orderedWorktreeIds: ['wt-a', 'wt-b', 'wt-c', 'wt-d', 'wt-e', WATCHED_WORKTREE], - activeWorktreeId: 'wt-active', - isRetained: () => true, - holdsLiveGuests: () => true, - isEvictable: (worktreeId) => - worktreeId === WATCHED_WORKTREE ? !browserTabsVetoGuestEviction(tabs) : true - }) -} - -describe('browser guest eviction veto', () => { - beforeEach(() => { - hydrateBrowserDrivers([]) - hydrateBrowserRemoteViewerPages([]) - }) - - // Parking preserves downloads; eviction unregisters the guest and main cancels them as on tab close. - it('spares a page that is still writing a download', () => { - let emitDownloadRequested: (event: { - downloadId: string - browserPageId: string - }) => void = () => {} - const noop = (): void => {} - vi.stubGlobal('window', { - api: { - browser: { - onDownloadRequested: (callback: typeof emitDownloadRequested) => { - emitDownloadRequested = callback - return noop - }, - onDownloadProgress: () => noop, - onDownloadFinished: () => noop - } - } - }) - const stopDownloadTracking = installBrowserPageDownloadActivityTracking() - try { - expect(browserTabsVetoGuestEviction(tabsFor(WATCHED_PAGE))).toBe(false) - expect(evictionRun(tabsFor(WATCHED_PAGE))).toContain(WATCHED_WORKTREE) - - emitDownloadRequested({ downloadId: 'dl-1', browserPageId: WATCHED_PAGE }) - expect(browserTabsVetoGuestEviction(tabsFor(WATCHED_PAGE))).toBe(true) - expect(evictionRun(tabsFor(WATCHED_PAGE))).not.toContain(WATCHED_WORKTREE) - } finally { - stopDownloadTracking() - vi.unstubAllGlobals() - } - }) - - it('holds the veto for a viewer on a non-active page of the same tab', () => { - const tabs = [ - { id: 'tab-1', activePageId: 'page-front', pageIds: ['page-front', WATCHED_PAGE] } - ] as unknown as BrowserWorkspace[] - hydrateBrowserRemoteViewerPages([WATCHED_PAGE]) - expect(evictionRun(tabs)).not.toContain(WATCHED_WORKTREE) - }) -}) diff --git a/src/renderer/src/components/browser-pane/host-guest/browser-guest-paint-retention.ts b/src/renderer/src/components/browser-pane/host-guest/browser-guest-paint-retention.ts index add09f275c1..0dc11766044 100644 --- a/src/renderer/src/components/browser-pane/host-guest/browser-guest-paint-retention.ts +++ b/src/renderer/src/components/browser-pane/host-guest/browser-guest-paint-retention.ts @@ -5,12 +5,12 @@ import { isBrowserPageMobileDriven, onBrowserDriverChange, useBrowserMobileDriverForAny -} from '../../../lib/pane-manager/browser-mobile-driver-state' +} from '../../../runtime/browser/ownership/browser-mobile-driver-state' +import { useBrowserRemoteViewerForAny } from '../../../runtime/browser/ownership/browser-remote-viewer-state' import { isBrowserPageRemotelyViewed, - onBrowserRemoteViewerChange, - useBrowserRemoteViewerForAny -} from '../../../lib/pane-manager/browser-remote-viewer-state' + onBrowserRemoteViewerChange +} from '../../../runtime/browser/ownership/browser-remote-viewer-store' import { isBrowserAutomationVisible, onBrowserAutomationVisibilityChange, diff --git a/src/renderer/src/components/right-sidebar/checks-panel/refresh/use-checks-panel-git-status-effects.tsx b/src/renderer/src/components/right-sidebar/checks-panel/refresh/use-checks-panel-git-status-effects.tsx index 9dc96a0fbeb..3591d5b9e76 100644 --- a/src/renderer/src/components/right-sidebar/checks-panel/refresh/use-checks-panel-git-status-effects.tsx +++ b/src/renderer/src/components/right-sidebar/checks-panel/refresh/use-checks-panel-git-status-effects.tsx @@ -11,7 +11,7 @@ import { import { createCoalescedPollRunner, type CoalescedPollRunner -} from '../../../../lib/repository-refresh/coalesced-poll-runner' +} from '../../../../lib/source-control/refresh/coalesced-poll-runner' import type { ChecksPanelControllerState } from '../use-checks-panel-controller-state' import type { ChecksPanelReviewState } from '../use-checks-panel-review-state' import type { ChecksPanelContextState } from '../use-checks-panel-context-state' diff --git a/src/renderer/src/components/right-sidebar/source-control/sync/git-status-refresh-scheduler.ts b/src/renderer/src/components/right-sidebar/source-control/sync/git-status-refresh-scheduler.ts index 7aec4737eae..106f981cad9 100644 --- a/src/renderer/src/components/right-sidebar/source-control/sync/git-status-refresh-scheduler.ts +++ b/src/renderer/src/components/right-sidebar/source-control/sync/git-status-refresh-scheduler.ts @@ -1,7 +1,7 @@ import { slowTaskRequiredIdleMs, type SlowTaskBackoffOptions -} from '../../../../lib/repository-refresh/coalesced-poll-runner' +} from '../../../../lib/source-control/refresh/coalesced-poll-runner' export type GitStatusRefreshReason = 'activity' | 'safety' diff --git a/src/renderer/src/components/right-sidebar/source-control/sync/stale-conflict-operation-poll.ts b/src/renderer/src/components/right-sidebar/source-control/sync/stale-conflict-operation-poll.ts index 972c3c36600..00a4d5f8d65 100644 --- a/src/renderer/src/components/right-sidebar/source-control/sync/stale-conflict-operation-poll.ts +++ b/src/renderer/src/components/right-sidebar/source-control/sync/stale-conflict-operation-poll.ts @@ -7,7 +7,7 @@ import { getRuntimeGitConflictOperation } from '@/runtime/source-control/git/run import { createCoalescedPollRunner, type SlowTaskBackoffOptions -} from '../../../../lib/repository-refresh/coalesced-poll-runner' +} from '../../../../lib/source-control/refresh/coalesced-poll-runner' import { installWindowVisibilityInterval, isWindowVisible diff --git a/src/renderer/src/components/right-sidebar/source-control/sync/use-branch-compare.ts b/src/renderer/src/components/right-sidebar/source-control/sync/use-branch-compare.ts index fcd1651944c..152e88e8baa 100644 --- a/src/renderer/src/components/right-sidebar/source-control/sync/use-branch-compare.ts +++ b/src/renderer/src/components/right-sidebar/source-control/sync/use-branch-compare.ts @@ -8,7 +8,7 @@ import { useAppStore } from '@/store/index' import { createLoadingBranchCompareSummary } from '@/store/slices/editor/git/branch-compare-state' import type { GitUpstreamStatus } from '../../../../../../shared/git/status/git-status-types' import { shouldClearBranchCompareForMissingBase } from '../../../../../../shared/source-control/decisions/source-control-base-ref-resolution' -import { slowTaskRequiredIdleMs } from '../../../../lib/repository-refresh/coalesced-poll-runner' +import { slowTaskRequiredIdleMs } from '../../../../lib/source-control/refresh/coalesced-poll-runner' import { useBranchCompareRefreshTriggers } from './use-branch-compare-refresh-triggers' // Slow polls idle for their own duration; explicit mutation/base/manual refreshes remain immediate. diff --git a/src/renderer/src/lib/sidebar-shadow/desktop-facts-shadow.ts b/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-desktop-facts.test-fixture.ts similarity index 89% rename from src/renderer/src/lib/sidebar-shadow/desktop-facts-shadow.ts rename to src/renderer/src/components/sidebar/worktree-list/parity/sidebar-desktop-facts.test-fixture.ts index 04df1fe809b..ec492721ebb 100644 --- a/src/renderer/src/lib/sidebar-shadow/desktop-facts-shadow.ts +++ b/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-desktop-facts.test-fixture.ts @@ -9,9 +9,9 @@ import { parseExecutionHostId, toRuntimeExecutionHostId, toSshExecutionHostId -} from '../../../../shared/execution-host' -import type { DesktopSidebarFacts } from '../../../../shared/sidebar-list/desktop-sidebar-facts' -import { folderWorkspaceKey } from '../../../../shared/workspace-scope' +} from '../../../../../../shared/execution-host' +import type { DesktopSidebarFacts } from '../../../../../../shared/sidebar-list/desktop-sidebar-facts' +import { folderWorkspaceKey } from '../../../../../../shared/workspace-scope' type WindowFactsState = Pick< AppState, diff --git a/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-list-parity.test-fixture.ts b/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-list-parity.test-fixture.ts index 99763960804..6527995cac8 100644 --- a/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-list-parity.test-fixture.ts +++ b/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-list-parity.test-fixture.ts @@ -33,12 +33,12 @@ import { sidebarWorkspaceRowFilters } from '../controls/sidebar-filter-actions' import { readWindowWorkspaceActivityFacts, toLocallyOwnedRows -} from '@/lib/sidebar-shadow/workspace-activity-shadow' -import { readWindowDesktopSidebarFacts } from '@/lib/sidebar-shadow/desktop-facts-shadow' +} from '@/components/sidebar/worktree-list/parity/sidebar-workspace-activity.test-fixture' +import { readWindowDesktopSidebarFacts } from '@/components/sidebar/worktree-list/parity/sidebar-desktop-facts.test-fixture' import { readWindowPairedServerHosts, readWindowPairedServerRows -} from '@/lib/sidebar-shadow/paired-server-rows-shadow' +} from '@/components/sidebar/worktree-list/parity/sidebar-paired-server-rows.test-fixture' import type { SidebarInputsFeedDomain } from '../../../../../../main/runtime/sidebar-inputs/sidebar-inputs-feed' import type { SidebarFeedDomains } from '../../../../../../shared/sidebar-list/sidebar-feed-domains' import { buildSidebarFeedCatalog } from '../../../../../../shared/sidebar-list/sidebar-feed-catalog' diff --git a/src/renderer/src/lib/sidebar-shadow/paired-server-rows-shadow.ts b/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-paired-server-rows.test-fixture.ts similarity index 76% rename from src/renderer/src/lib/sidebar-shadow/paired-server-rows-shadow.ts rename to src/renderer/src/components/sidebar/worktree-list/parity/sidebar-paired-server-rows.test-fixture.ts index c0fc93244e4..53baa7dc491 100644 --- a/src/renderer/src/lib/sidebar-shadow/paired-server-rows-shadow.ts +++ b/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-paired-server-rows.test-fixture.ts @@ -5,22 +5,22 @@ import { import type { AppState } from '@/store/types' import { getIndexedAllWorktrees } from '@/store/slices/worktrees/listing/worktree-catalog-index' import { getIndexedRepoMap } from '@/store/repos/repository-catalog-index' -import { projectResolvedWorktreeLineage } from '../../../../shared/resolved-worktree-lineage' -import type { WorktreeLineage } from '../../../../shared/worktree/lineage-types' -import { folderWorkspaceToWorktree } from '../../../../shared/folder-workspaces/folder-workspace-worktree' +import { projectResolvedWorktreeLineage } from '../../../../../../shared/resolved-worktree-lineage' +import type { WorktreeLineage } from '../../../../../../shared/worktree/lineage-types' +import { folderWorkspaceToWorktree } from '../../../../../../shared/folder-workspaces/folder-workspace-worktree' import { pairedServerRowOfWorktree, type PairedServerHostFacts, type PairedServerWorkspaceRow -} from '../../../../shared/sidebar-list/paired-server-rows' -import { describePairedServerHosts } from '../../../../shared/remote-runtime/host-catalog/paired-server-hosts' -import { getHostDisplayLabelOverrides } from '../../../../shared/host-setting-overrides' -import { getWorktreeHostIdentity } from '../../../../shared/worktree/host-qualified-identity' -import type { Worktree } from '../../../../shared/worktree/types' -import { lineageStoreKeyOfRow } from '../../../../shared/worktree/cross-store-lineage-resolution' -import { validCrossStoreLineage } from '../../../../shared/worktree/cross-store-lineage-records' -import { worktreeWorkspaceKey } from '../../../../shared/workspace-scope' -import { folderLineageEdgeOf } from '../../../../shared/worktree/folder-lineage-parents' +} from '../../../../../../shared/sidebar-list/paired-server-rows' +import { describePairedServerHosts } from '../../../../../../shared/remote-runtime/host-catalog/paired-server-hosts' +import { getHostDisplayLabelOverrides } from '../../../../../../shared/host-setting-overrides' +import { getWorktreeHostIdentity } from '../../../../../../shared/worktree/host-qualified-identity' +import type { Worktree } from '../../../../../../shared/worktree/types' +import { lineageStoreKeyOfRow } from '../../../../../../shared/worktree/cross-store-lineage-resolution' +import { validCrossStoreLineage } from '../../../../../../shared/worktree/cross-store-lineage-records' +import { worktreeWorkspaceKey } from '../../../../../../shared/workspace-scope' +import { folderLineageEdgeOf } from '../../../../../../shared/worktree/folder-lineage-parents' /** The paired-server rows the window fetched itself, keyed by host-qualified id, for these hosts. */ export function readWindowPairedServerRows( diff --git a/src/renderer/src/lib/sidebar-shadow/workspace-activity-shadow.ts b/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-workspace-activity.test-fixture.ts similarity index 92% rename from src/renderer/src/lib/sidebar-shadow/workspace-activity-shadow.ts rename to src/renderer/src/components/sidebar/worktree-list/parity/sidebar-workspace-activity.test-fixture.ts index 550945104b4..bbbeaa6f4de 100644 --- a/src/renderer/src/lib/sidebar-shadow/workspace-activity-shadow.ts +++ b/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-workspace-activity.test-fixture.ts @@ -6,18 +6,18 @@ import type { AppState } from '@/store/types' import { getIndexedAllWorktrees } from '@/store/slices/worktrees/listing/worktree-catalog-index' import { getIndexedRepoMap } from '@/store/repos/repository-catalog-index' import { getLiveAgentStatusByWorktreeId } from '@/lib/workspace/activity/worktree-activity-state' -import { tabHasLivePty } from '../../../../shared/tab-has-live-pty' +import { tabHasLivePty } from '../../../../../../shared/tab-has-live-pty' import { getWorktreeIdsWithStructuredChat } from '@/components/sidebar/worktree-list/listing/visible-worktree-activity-inputs' import { getRepoExecutionHostId, parseExecutionHostId, type ExecutionHostId -} from '../../../../shared/execution-host' -import { folderWorkspaceKey } from '../../../../shared/workspace-scope' +} from '../../../../../../shared/execution-host' +import { folderWorkspaceKey } from '../../../../../../shared/workspace-scope' import { isIdleWorkspaceActivity, type WorkspaceActivityFacts -} from '../../../../shared/sidebar-list/workspace-activity-facts' +} from '../../../../../../shared/sidebar-list/workspace-activity-facts' /** The window's own activity facts, from the same store slices the sidebar filter reads. */ export function readWindowWorkspaceActivityFacts( diff --git a/src/renderer/src/lib/sidebar-shadow/workspace-activity-shadow.test.ts b/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-workspace-activity.test.ts similarity index 88% rename from src/renderer/src/lib/sidebar-shadow/workspace-activity-shadow.test.ts rename to src/renderer/src/components/sidebar/worktree-list/parity/sidebar-workspace-activity.test.ts index 3bcedc91e8c..1a4f12fb30e 100644 --- a/src/renderer/src/lib/sidebar-shadow/workspace-activity-shadow.test.ts +++ b/src/renderer/src/components/sidebar/worktree-list/parity/sidebar-workspace-activity.test.ts @@ -1,8 +1,8 @@ import { describe, expect, it } from 'vitest' -import { useAppStore } from '@/store' +import { useAppStore } from '@/store/index' import { makeWorktree } from '@/components/worktree-jump/worktree-jump-palette-test-fixtures' -import { toLocallyOwnedRows } from './workspace-activity-shadow' +import { toLocallyOwnedRows } from './sidebar-workspace-activity.test-fixture' describe('toLocallyOwnedRows', () => { it('leaves out an id a paired server also lists, since bare-id activity names no one host', () => { diff --git a/src/renderer/src/components/tab-bar/ClientHostedBrowserTabRows.tsx b/src/renderer/src/components/tab-bar/ClientHostedBrowserTabRows.tsx index 74cdf7bd6e7..1a6abf80ff9 100644 --- a/src/renderer/src/components/tab-bar/ClientHostedBrowserTabRows.tsx +++ b/src/renderer/src/components/tab-bar/ClientHostedBrowserTabRows.tsx @@ -6,7 +6,7 @@ import type { ClientHostedBrowserRow } from '../../../../shared/client-hosted-br import { selectClientHostedBrowserRow, useActiveClientHostedBrowserRowId -} from '@/lib/pane-manager/client-hosted-browser-row-state' +} from '@/runtime/browser/client-hosting/client-hosted-browser-row-state' import { closeClientHostedBrowserRow } from '../../runtime/browser/closing/client-hosted-browser-row-close' import ClientHostedBrowserTab from './ClientHostedBrowserTab' diff --git a/src/renderer/src/components/tab-bar/TabBar.tsx b/src/renderer/src/components/tab-bar/TabBar.tsx index 292f2f4b697..d84ef3c4952 100644 --- a/src/renderer/src/components/tab-bar/TabBar.tsx +++ b/src/renderer/src/components/tab-bar/TabBar.tsx @@ -9,7 +9,7 @@ import { usePluginNewFileMenuOptions } from './use-plugin-new-file-menu-options' import { useTabBarItemProjection } from './use-tab-bar-item-projection' import { renderTabBarSurface } from './tab-bar-surface' import { useTabBarItemActions } from './use-tab-bar-item-actions' -import { useActiveClientHostedBrowserRowId } from '@/lib/pane-manager/client-hosted-browser-row-state' +import { useActiveClientHostedBrowserRowId } from '@/runtime/browser/client-hosting/client-hosted-browser-row-state' import { useEditorGroupFileDropOwner } from '../editor/use-editor-group-file-drop-owner' import { useTranslate } from '@/i18n/use-translate' diff --git a/src/renderer/src/components/tab-bar/TabBarItemRow.test.tsx b/src/renderer/src/components/tab-bar/TabBarItemRow.test.tsx deleted file mode 100644 index 2be37adac80..00000000000 --- a/src/renderer/src/components/tab-bar/TabBarItemRow.test.tsx +++ /dev/null @@ -1,271 +0,0 @@ -// @vitest-environment happy-dom - -import { act, startTransition, Suspense, use } from 'react' -import { createRoot, type Root } from 'react-dom/client' -import { afterEach, describe, expect, it, vi } from 'vitest' -import { clearClientHostedBrowserRowSelection } from '@/lib/pane-manager/client-hosted-browser-row-state' - -import type { TabBarItem } from './tab-bar-item-model' -import type { WorkspaceVisibleTabType } from '../../../../shared/tab-types' -import type { AgentType } from '../../../../shared/agent-status/contracts/agent-status-types' -import { - renderTabBarItems, - type TabBarItemSurfaceProps, - type TabBarItemSurfaceRuntime -} from './tab-bar-item-surface' -import { useTabBarItemActions } from './use-tab-bar-item-actions' - -globalThis.IS_REACT_ACT_ENVIRONMENT = true - -type TabProps = { - tab?: { id: string; title: string } - agent?: string | null - nativeChat?: { sessionId: string; agent: string | undefined } - isActive: boolean - isPinned: boolean - onActivate: (id: string) => void - onDuplicate?: () => void - gitStatus?: string | null -} - -const tabRenders: { id: string; props: TabProps }[] = [] - -vi.mock('./SortableTab', () => ({ - default: (props: TabProps) => { - tabRenders.push({ id: props.tab!.id, props }) - return null - } -})) -vi.mock('./BrowserTab', () => ({ - default: (props: TabProps) => { - tabRenders.push({ id: props.tab!.id, props }) - return null - }, - getBrowserTabLabel: () => '' -})) -vi.mock('./EditorFileTab', () => ({ - default: (props: TabProps & { file: { id: string } }) => { - tabRenders.push({ id: props.file.id, props }) - return null - } -})) - -/** Rebuilt on every call, like the strip's projections: equal content, never the same objects. */ -function buildItems({ - terminalPinned = false, - browserTitle = 'Example', - sessionLabel = 'Session', - sessionAgent -}: { - terminalPinned?: boolean - browserTitle?: string - sessionLabel?: string - sessionAgent?: AgentType -} = {}): TabBarItem[] { - return [ - { - type: 'terminal', - id: 'terminal-1', - unifiedTabId: 'unified-terminal-1', - isPinned: terminalPinned, - data: { - id: 'terminal-1', - ptyId: null, - worktreeId: 'wt-1', - title: 'zsh', - generatedTitle: 'Fix the login bug', - customTitle: null, - color: null, - sortOrder: 0, - createdAt: 0 - } - }, - { - type: 'browser', - id: 'browser-1', - unifiedTabId: 'unified-browser-1', - isPinned: false, - data: { - id: 'browser-1', - worktreeId: 'wt-1', - url: 'https://example.test', - title: browserTitle, - loading: false, - faviconUrl: null, - canGoBack: false, - canGoForward: false, - loadError: null, - createdAt: 0 - } - }, - { - type: 'editor', - id: 'file-1', - unifiedTabId: 'unified-file-1', - isPinned: false, - data: { - id: 'file-1', - filePath: '/repo/notes.md', - relativePath: 'notes.md', - worktreeId: 'wt-1', - language: 'markdown', - isPreview: false, - isDirty: false, - mode: 'edit' - } - }, - { - type: 'agent-session', - id: 'session-1', - unifiedTabId: 'session-1', - isPinned: false, - data: { - id: 'session-1', - entityId: 'session-1', - groupId: 'group-1', - worktreeId: 'wt-1', - contentType: 'agent-session', - ...(sessionAgent ? { agentSessionAgent: sessionAgent } : {}), - label: sessionLabel, - customLabel: null, - color: null, - sortOrder: 3, - createdAt: 0 - } - } - ] -} - -type StripInputs = { - onActivate?: (id: string) => void - generatedTabTitlesEnabled?: boolean - managedBrowserCreationEnabled?: boolean - terminalPinned?: boolean - browserTitle?: string - sessionLabel?: string - sessionAgent?: AgentType - activeTabType?: WorkspaceVisibleTabType - activeClientHostedBrowserRowId?: string | null - statusByRelativePath?: TabBarItemSurfaceRuntime['statusByRelativePath'] -} - -function Strip({ - onActivate = NOOP, - generatedTabTitlesEnabled = false, - managedBrowserCreationEnabled = false, - terminalPinned, - browserTitle, - sessionLabel, - sessionAgent, - activeTabType = 'terminal', - activeClientHostedBrowserRowId = null, - statusByRelativePath = STATUS_BY_RELATIVE_PATH -}: StripInputs): React.JSX.Element { - const props: TabBarItemSurfaceProps = { - worktreeId: 'wt-1', - activeTabId: 'terminal-1', - activeFileId: 'file-1', - activeBrowserTabId: 'browser-1', - activeSimulatorTabId: null, - activeTabType, - expandedPaneByTabId: {} - } - const runtime: TabBarItemSurfaceRuntime = { - resolvedGroupId: 'group-1', - generatedTabTitlesEnabled, - unifiedTabByVisibleId: new Map(), - tabAgentTypesByTabId: {}, - nativeChatTabWideFallbackUnsafeTabsById: {}, - nativeChatTranscriptIsLocalReadable: false, - managedBrowserCreationEnabled, - statusByRelativePath - } - const actions = useTabBarItemActions({ - props: { - worktreeId: 'wt-1', - onActivate, - onActivateFile: onActivate, - onActivateBrowserTab: onActivate, - onActivateAgentSession: onActivate, - onClose: NOOP, - onCloseOthers: NOOP, - onCloseToRight: NOOP, - onCloseToLeft: NOOP, - onSetCustomTitle: NOOP, - onSetTabColor: NOOP, - onTogglePaneExpand: NOOP - }, - togglePinned: NOOP, - toggleTabViewMode: NOOP - }) - return ( - <> - {renderTabBarItems({ - items: buildItems({ terminalPinned, browserTitle, sessionLabel, sessionAgent }), - props, - runtime, - actions, - dropIndicatorByVisibleId: new Map(), - includeTopTabBorder: true, - activeClientHostedBrowserRowId - })} - - ) -} - -const NOOP = (): void => {} -const STATUS_BY_RELATIVE_PATH: TabBarItemSurfaceRuntime['statusByRelativePath'] = new Map() - -let root: Root | null = null - -const NEVER_RESOLVES = new Promise(() => {}) - -function Suspended(): null { - use(NEVER_RESOLVES) - return null -} - -function stripTree(inputs: StripInputs, suspended = false): React.JSX.Element { - return ( - - - {suspended ? : null} - - ) -} - -function renderStrip(inputs: StripInputs = {}): void { - if (!root) { - root = createRoot(document.createElement('div')) - } - act(() => root!.render(stripTree(inputs))) -} - -function lastRender(tabId: string): TabProps { - return tabRenders.findLast((render) => render.id === tabId)!.props -} - -afterEach(() => { - act(() => root?.unmount()) - root = null - tabRenders.length = 0 - clearClientHostedBrowserRowSelection() -}) - -describe('tab strip rows', () => { - it('keeps the committed handler when React abandons a render', async () => { - const committed = vi.fn() - const abandoned = vi.fn() - renderStrip({ onActivate: committed }) - await act(async () => - startTransition(() => root!.render(stripTree({ onActivate: abandoned }, true))) - ) - - lastRender('terminal-1').onActivate('terminal-1') - - expect(committed).toHaveBeenCalledWith('terminal-1') - expect(abandoned).not.toHaveBeenCalled() - }) -}) - -/** Client-hosted rows overlay panes without changing activeTabId; suppress real-tab underlines while they are selected. */ diff --git a/src/renderer/src/components/tab-bar/use-tab-bar-item-actions.ts b/src/renderer/src/components/tab-bar/use-tab-bar-item-actions.ts index c6454250f75..f4d0a88bfb0 100644 --- a/src/renderer/src/components/tab-bar/use-tab-bar-item-actions.ts +++ b/src/renderer/src/components/tab-bar/use-tab-bar-item-actions.ts @@ -1,5 +1,5 @@ import { useLayoutEffect, useMemo, useRef } from 'react' -import { clearClientHostedBrowserRowSelection } from '@/lib/pane-manager/client-hosted-browser-row-state' +import { clearClientHostedBrowserRowSelection } from '@/runtime/browser/client-hosting/client-hosted-browser-row-state' import { noteUserWorkspaceSwitch } from '@/lib/workspace/activation/workspace-switch-user-act' import { useAppStore } from '@/store' import { ORCA_LAYOUT_OWNER_IS_MAIN } from '../../../../shared/layout-owner-switch' diff --git a/src/renderer/src/components/tab-group/TabGroupPanel.os-file-drop.test.tsx b/src/renderer/src/components/tab-group/TabGroupPanel.os-file-drop.test.tsx deleted file mode 100644 index 3b05b8aba3b..00000000000 --- a/src/renderer/src/components/tab-group/TabGroupPanel.os-file-drop.test.tsx +++ /dev/null @@ -1,118 +0,0 @@ -// @vitest-environment happy-dom -import { useEffect, useRef } from 'react' -import { act, cleanup, render, waitFor } from '@testing-library/react' -import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest' - -const mocks = vi.hoisted(() => { - const state = { - activeWorktreeId: 'wt-active', - rightSidebarOpen: true, - sidebarOpen: true, - groupsByWorktree: {} - } - const useAppStore = (selector: (current: typeof state) => unknown): unknown => selector(state) - useAppStore.getState = () => state - return { capture: vi.fn(), deliver: vi.fn(), monacoSawDrop: vi.fn(), useAppStore } -}) -vi.mock('@/store', () => ({ useAppStore: mocks.useAppStore })) -vi.mock('../../store', () => ({ useAppStore: mocks.useAppStore })) -vi.mock('../editor/editor-dropped-file-open', () => ({ - captureEditorFileDropOpen: mocks.capture, - editorGroupStillExists: () => true -})) -vi.mock('../tab-bar/TabBar', () => ({ default: () => null })) -vi.mock('../tab-bar/TabBarQuickCommandsButton', () => ({ TabBarQuickCommandsButton: () => null })) -vi.mock('@/lib/pane-manager/client-hosted-browser-row-state', () => ({ - useClientHostedBrowserRows: () => [] -})) -vi.mock('./useTabGroupWorkspaceModel', () => ({ - useTabGroupWorkspaceModel: () => ({ - activeTab: { id: 'tab-1', entityId: 'file-1', contentType: 'editor' }, - agentSessionItems: [], - browserItems: [], - commands: { focusGroup: () => undefined }, - editorItems: [], - tabBarOrder: [], - terminalTabs: [], - groupTabs: [], - expandedPaneByTabId: {} - }) -})) -// Stands in for Monaco, which listens for drops on its own DOM node. -vi.mock('../editor/EditorPanel', () => ({ - default: function FakeMonaco() { - const ref = useRef(null) - useEffect(() => { - const node = ref.current - node?.addEventListener('drop', mocks.monacoSawDrop) - node?.addEventListener('dragover', mocks.monacoSawDrop) - return () => { - node?.removeEventListener('drop', mocks.monacoSawDrop) - node?.removeEventListener('dragover', mocks.monacoSawDrop) - } - }, []) - return ( -
-