mirror of
https://github.com/stablyai/orca.git
synced 2026-11-01 01:27:35 +00:00
2fc2f92a386471ccfbe8277cea3df71d2f6ec0cf
19
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
  Neilandstevelliu
|
347d0e46a1 |
feat(cli): set the title of one terminal pane (#30485)
* feat(cli): set the title of one terminal pane Adds `orca terminal set-pane-title --terminal <handle> --title <text>`; `--title ""` clears it. The runtime resolves the live handle to its saved leaf and asks the desktop window to name that pane through a new optional notifier hook (`ui:setPaneTitle`). Mounted panes apply it through the same local setter the GUI uses; parked panes get it in the saved layout their next mount reads. The tab's own name is never touched. Rebuilt on current main from #24755. Co-authored-by: stevelliu <stevelliu@tencent.com> * fix(renderer): a main-owned layout takes a CLI pane title before the window shows it * fix(renderer): a CLI pane title reaches a never-mounted tab with a rootless layout --------- Co-authored-by: stevelliu <stevelliu@tencent.com> |
||
 
|
4940654793 |
fix(cli): --host and --project accept the names host list and project list print (#30438)
* Accept host aliases and infer the host for a single-setup project `orca worktree create --project scalp-it` failed with "Project is not set up on the selected host" even though the project has exactly one ready setup, and `--host trader-local` was rejected as an invalid value even though `orca host list` prints that name. Three separate causes: - `--project` was matched against the stored `projectId` only, which is a provider-scoped string (`github:owner/repo`). The display name shown by `orca project list` never matched, so the lookup found nothing and the failure was reported as a missing host. - With no `--host`, the host fallback existed but was unreachable, because the project lookup above had already filtered to zero candidates. - `parseExecutionHostId` accepts only `local`, `ssh:<id>` and `runtime:<id>`, so a bare alias could not parse at all. `ssh:<label>` already resolved labels; only the unprefixed spelling was missing. Now: - `--project` takes the project id or the display name, resolving through `project.list` and falling back to the setup's own display name for hosts that predate that method. Two projects sharing a name are refused rather than guessed. - `--host` is only required when the project is set up on more than one host. A single setup is used directly; an ambiguous one fails with every host and the flag to select it. - `--host <alias>` resolves against both axes: paired servers by name before routing, SSH targets by label once a client exists. An alias that names a machine on both is refused with the prefixed flag for each. Errors now echo the flag the caller typed instead of an `ssh:` spelling they did not use, and carry the hosts to retry with in `nextSteps`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Refuse a host alias that collides with duplicated SSH labels assertAliasNamesOneHost asked findSshTargetByName whether the alias also named an SSH target, but that function returns undefined for a label two targets share — deliberately, so an ambiguous label never resolves. Read as a boolean it says "no SSH target by that name", so an alias naming one paired server and two same-labelled SSH targets looked conflict-free and resolved to the server. The opposite direction was already covered: a unique SSH target against two servers sharing a name fails through assertEnvironmentNameUnambiguous. Fall back to ambiguousSshTargets so both axes refuse, and list every colliding target in nextSteps rather than only the first. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Only fall back to setup display names on a server without project.list findProjectsByName caught every error from project.list and returned an empty list, so a network failure, a permission error, or a genuine server fault was indistinguishable from "no project by that name" — and the caller was then told `Project is not set up on the selected host` for a project that exists and is set up. Narrow the catch to method_not_found, the only code that means the weaker fallback is warranted, matching callProjectHostSetup and listSshTargets in the same layer. The function now returns null for that version gap and an empty array for a real miss, so the two cases stay distinguishable at the call site. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Refuse a legacy display-name match that spans two projects The pre-project.list fallback matches ProjectHostSetup.displayName, which is the repo's folder name rather than the project's name, so two projects holding repos with the same folder name both match. selectSetupOnHost would then have taken the first candidate — targeting a project the caller never named. Reject when the matched setups span more than one projectId and name each one, the same rule the project.list path already applies to duplicate display names. A single project with several setups is unaffected; the host selection below still decides between them. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Spell the --host placeholder as <host> in the project setup usage lines The worktree create usage was updated to <host> when aliases became valid, but the four project setup-* lines kept <host-id> in both the command specs and the root help text they mirror. Those commands accept an alias too, so the narrower placeholder hid it from anyone reading help. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Document host names and the optional --host in the CLI docs The published CLI reference described `--host` as taking only `local`, `ssh:<target-id>` or `runtime:<environment-id>`, and the automations page showed `--host <hostId>` as required alongside `--project`. Both now understate what the commands accept, which is how the original report started: `orca host list` prints a name, the docs never said you could type it, so the generated id looked mandatory. Spell out that `--host` takes the bare name from `host list` on either axis, that a name resolves only when unique and is otherwise refused with the prefixed flag for each candidate, and that `--host` is unnecessary when the project has a single setup. Also note `--project` accepts the display name from `orca project list`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Name the runtime host id as the paired-server tie-breaker in the docs The host-selection page told readers to fall back to ids when a name is ambiguous but only ever showed `--environment <server-name>` for paired servers — and that selector is exactly the one that cannot separate two servers sharing a name. Spell out the pair the resolver actually emits in its nextSteps: `--host ssh:<target-id>` and `--host runtime:<environment-id>`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> * Spell the automations create usage as --project <id-or-name> [--host <host>] `automations create` resolves --project and --host through the same alias-capable resolver as `worktree create`, and its own notes already say --project takes a display name and --host takes a bare host name. The usage line still read `--project <id> [--host <id>]`, the last spot left on the old spelling (flagged by CodeRabbit after the main merge). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * Give the fake runtime client cast in the host alias tests its SAFETY note Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(cli): accept every name host list and project list print, everywhere --project/--host is taken Extends the alias work to the commands it did not reach and the names it missed: - `project setups --project <display name>` filtered by projectId only. - `project setup-existing-folder|setup-clone|setup-create --project <name>` now send the project id that name resolves to (pass-through for ids and for servers without project.list). - `--host "this machine"` (the local row's name) resolves to local, and is refused if an SSH target or paired server shares it. - One pair of wrapping quotes is ignored, since cmd.exe passes single quotes through literally. - With --environment/--pairing-code, a bare --host name is a machine the routed server reaches, not a different locally paired server. - An unknown project is reported as a project miss, not a host miss. Co-authored-by: Younghwan Chae <chyohw97@gmail.com> * fix(cli): keep exact project ids over names, and refuse an unverifiable host name An exact --project id whose only setup is not ready no longer falls through to a project whose display name equals that id. A bare --host name that matches a paired server now routes there only once this machine's SSH targets were actually read, and under --environment/--pairing-code a bare name always means a machine the routed server reaches. Co-authored-by: Younghwan Chae <chyohw97@gmail.com> --------- Co-authored-by: younghwan <chyohw97@gmail.com> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
 Jinjing
|
232f6048e7 | Remove obsolete Hermes cron documentation and translations (#28877) | ||
 NeilandNikita Maketryuk
|
b428def8f1 |
feat(cli): open files at a requested line and column (#28240)
Accept positive safe-integer line and column flags through the existing file-open path. Reject invalid or incomplete positions before changing the editor. Bind deferred reveals to the requested workspace and tab, including floating panels, and clamp oversized positions without moving another editor or stealing late focus. Co-authored-by: Nikita Maketryuk <n.maketryuk@lince.studio> |
||
  OrcaWinandOrca Worker
|
83056e5edb |
fix(skills): judge official skill updates by ownership, not snapshot matching (#27603)
* fix(skills): judge official skill updates by ownership, not snapshot matching Orca decided whether an installed official skill was pristine by matching its bytes against a shipped list of past snapshots, falling back to the skills CLI's lock hash. Copies installed between releases, CRLF copies, and installs whose lock recorded a sha256 hash all read "doesn't match the official version" and could not be updated (#26841), and older content could be labelled newer and never offered. Now an official skill is Orca's when the skills.sh global lock says it came from stablyai/orca, and it is current only when its listed files equal the bundled manifest (line endings and sidecars ignored); anything else is outdated. Install and update are one pinned `skills add .../tree/v<appVersion>` command, used by the in-app runner, `orca skills update`, and the Settings rails. The update dialog preselects Orca-owned rows and lists same-named skills from elsewhere unchecked. Removes the snapshot registry, release mapping, git-tree identity, lock trust and convergence logic, and the release-cut provenance step. * ci(skills): run the pinned-add round trip on a skills CLI that can fetch a commit * fix(skills): let orca skills update target the shared folder on a bare host Update became the same pinned add as install, so it inherited install's refusal when no agent is detected. That broke the bare-host update the old command guaranteed, and the built-CLI check (update --dry-run, no --agent) on CI runners. The in-app runner already fell back to universal; both now share one detectSkillsCliUpdateAgentKeys rule. Also brings the CLI skills docs off `npx skills update`. * refactor(skills): derive the update selection from the grouped rows The selection hook took the live and the retained inventory, re-derived Orca ownership from installations beside the grouping's own orcaOwned, and kept a separate display count. Groups already carry both eligibility and ownership, so the hook reads them once; the button's existing live-scan gate is the only authorization check. The unselected line is inlined into the headline. * refactor(skills): drop helpers the ownership design made redundant - isOwnerManagedSkillScope only negated into the participation check once plugin caches stopped participating; fold it in. - Eligibility re-checked resolvedPath/physicalIdentity, which observation already turns into 'inaccessible'. - The manifest verifier probed access() before a readFile that throws anyway. - orchestration-install-command.ts only forwarded one constant from the lib module its three importers already import. * i18n(skills): stop naming npx skills update in four skipped-copy strings The es/ja/ko/zh cantUpdateReason and tipDuplicate strings still named `npx skills update`, which Orca no longer runs for its skills. Say 'the update command' instead. * refactor(skills): inline the single-use pinned source URL helper orcaSkillsSourceUrl was exported but only called by the builder beside it. * fix(skills): let an update choice lapse when the install source changes The selection froze every checked name, Orca's preselection included, the moment the user ticked anything. A re-check that then attributed a name to another source kept it checked, replacing it without consent. Store only the user's explicit choices, keyed by name and lock source, and default to ownership otherwise. * fix(skills): label the empty update action 'Update skills', not 'Update 0 skills' The button stays whenever a row is updatable, so unticking every row (or a dialog with only non-Orca copies outdated) doesn't reflow the footer. At zero selected it stays disabled and reads 'Update skills' instead of a zero count. The label lives beside the dialog's other count copy. * fix(skills): let keyboard focus open the Not from Orca tooltip The source of a non-Orca copy is only in the tooltip, and its trigger was a plain span. Make it focusable like Orca's other tooltip-on-text triggers (tabIndex 0, cursor-help, the row primitive's focus ring), as in the agreed mock. * test(skills): hold the round-trip add equal to the shared install builder The CI round trip hardcoded its skills add argv, so it could drift from what Orca runs. A .mjs script can't import the TS builder (extensionless imports), so the argv moves into an exported roundTripAddArgs and a unit test asserts it is the builder's unattended add with the tag swapped for the PR commit. The script runs under the same main guard as the manifest generator. * i18n(skills): translate the skill row's location count and duplicate reason es, ko and zh still carried English for SkillUpdateRow.oneLocation and manyLocations, and es and ko for SkillFreshnessRow.skippedReasonDuplicate. * fix(skills): pin official skill installs to the build commit and skills CLI `skills add .../tree/v<version>` only works for stable and RC builds: hourly, daily and adhoc builds publish elsewhere, so stablyai/orca has no tag for their version and both install and update failed there. Every build is made from a commit that exists in stablyai/orca, so the build now stamps `git rev-parse HEAD` and the one shared builder pins to `tree/<commit>`, which equals the bundled skills by construction. Main, renderer, web and vitest get it as ORCA_BUILD_COMMIT (replacing ORCA_APP_VERSION); the tsc-built CLI reads it from the compiled out/package.json boundary. A missing or malformed commit fails the build or the command with a clear error instead of building a bad URL. The builder also pins the skills CLI (`npx skills@1.7.2`): 1.5.x cannot fetch a commit, and a bare `npx skills` can reuse a stale cache or resolve to another binary (#18687). The CI round-trip script reuses the same constant. * ci(skills): run the round trip on the shared pinned skills CLI version * docs(skills): document the commit-pinned skills add * refactor(skills): keep the pinned skills CLI version in the install builder skills-cli-version.ts existed only so the plain-Node round-trip script could strip-load a .ts file, which needed --disable-warning in CI. The version now lives beside the builder that uses it; the script keeps a literal, and the existing round-trip test already asserts its argv, skills@<version> included, equals the builder's. * refactor(skills): validate the build commit once, where it is resolved resolveBuildSourceCommit already refuses anything but a full hash, and both the define and the CLI's stamped package.json come from it. The builder's second check could only fire as a module-load throw in the renderer's command constants. Also drops the resolver's unused cwd parameter. * test: stamp a fixed build commit instead of running git in the vitest config Tests only interpolate ORCA_BUILD_COMMIT, so they need not read HEAD (or fail outside a git checkout) on every run. * refactor(settings): match only skills add in the Windows npx preflight Every official skill command is now the builder's skills@<version> add; no rail emits skills update any more. * fix(skills): fail an update that leaves a broken skill link behind Eligibility already skips names with a broken link before the run, so one afterwards is the run's own damage; report it instead of claiming success. * ci(release): refuse to tag when the skill manifest is stale The removed --release step implicitly checked the committed manifest against skills/. Keep that guarantee with a verify-only run; it writes nothing. --------- Co-authored-by: Orca Worker <orca-worker@localhost> |
||
|
Neil
|
15717a4e53 |
Add CLI commands to manage and find workspace references (#26807)
* Add CLI commands to manage and find workspace references * Fix reference review findings: source-context twins, legacy link collapse, closed chats, old-host errors - Keep same-URL links with different source contexts separate; CLI add/remove no longer drops a twin - Restore merging a bare legacy link into its single matching URL link (no duplicate PR rows, no resurrection on older-client removal) - Carry absorbed data through collapse chains; resolve repoId conflicts deterministically - Identity: ignore www., case-insensitive GitHub/GitLab/Gitea paths (stored URL keeps case), no Jira match on git hosts; find accepts self-hosted issue URLs - reference find: skip closed native chats, require a session id for linked, include an explicitly named archived workspace, warm-cache fallback for --worktree - Selector parity: collapse same-host duplicate paths, issue: ignores unlinked rows - CLI: clear update-host error for hosts without reference.*, local flag checks before host probe, help text fixes * Fix CI: add reference rm alias for verb policy; expect linkedItems in worktree ps summary * Simplify reference CLI: share identifier/legacy-match helpers, drop redundant dedupe |
||
|
Jinjing
|
55eb48c137 |
Let automations pass extra agent arguments on each run (#26659)
* Let automations pass extra agent arguments on each run Automations could only use the host's default agent Arguments, so a per-automation choice such as a different model needed a settings change that affected every run. Automations can now save extra arguments that are appended to the host's defaults on each fresh run. Extras are allowlisted to model, effort, and Claude's --add-dir, and are refused for permission, config, prompt, and session options. - Add the field through the CLI (`automations create/edit --extra-agent-args`), the RPC contract, the automation store, and the editor's Advanced section, with validation at save and at launch. - Merge extras over the defaults by replacing a model or effort option the defaults already set, then re-tokenize to confirm the launch argv matches what was validated. - Gate writes behind a new runtime capability, and refuse to save or migrate extras to a host that would silently drop them. - Require a fresh session, since reused sessions cannot take new arguments. * Simplify automation argument validation and launch handling * improve the extra args display * fix translation * Validate automation extra args during ORCA migration Dormant automations imported from the migration manifest skipped the extra-args check. A disallowed flag, or Reuse combined with extra args, could persist and then fail every later run. The manifest parser now rejects those rows. Also move the automation runtime capability constants into `automation-runtime-capabilities.ts`. `protocol-version.ts` re-exports them, so existing imports keep working. * add more demo with effort |
||
+2        
|
8e5080c132 |
Validate OpenCode worker model preferences on the execution host (#24624)
* feat(orchestration): support OpenCode worker model selection Allow supervised OpenCode workers to use per-launch model overrides through the existing launch-preference and receipt path. Preserve existing OpenCode agent arguments while replacing only model flags, and reject unsupported effort values explicitly. Keep Native Chat option exposure unchanged and update the worker CLI and orchestration guidance. * fix: wait for OpenCode worker composer before first dispatch Reuse captured composer readiness on local and paired execution hosts and revoke launching-shell paste anchors. Co-authored-by: Brennan Benson <79079362+brennanb2025@users.noreply.github.com> * feat(opencode): probe execution-host CLI capabilities * fix(opencode): select plugin default for execution host loader * fix(opencode): limit prompt prefill capability to verified release * feat(orchestration): gate OpenCode worker model preferences by host capability Co-authored-by: user141514 <user141514@users.noreply.github.com> * feat(opencode): probe launch capabilities on the execution host * feat(opencode): probe execution-host CLI capabilities * feat(opencode): probe launch capabilities on the execution host * feat(orchestration): resolve explicitly configured command aliases * fix(opencode): select plugin loader for the launched host binary * fix(opencode): match WSL probe cwd and declared guest environment * fix(orchestration): verify available OpenCode model on execution host * fix(opencode): preserve launch environment deletion boundaries * wip(opencode): authorize native startup prompt intent at execution owner * fix(opencode): atomically replace status plugin entrypoints * fix(opencode): retain plugin permissions across restrictive umasks * test(opencode): resolve permission fixture from primary cwd * feat(opencode): install startup prompt plugin independently of status hooks * fix(opencode): wait for admitted startup intent and preserve failed-launch briefs * fix(opencode): confine overlay manifest cleanup to owned directories Co-authored-by: Adnan Khan <adnank11427@gmail.com> * fix: wait for OpenCode worker composer before first dispatch Reuse captured composer readiness on local and paired execution hosts and revoke launching-shell paste anchors. Co-authored-by: Brennan Benson <79079362+brennanb2025@users.noreply.github.com> * feat(opencode): probe execution-host CLI capabilities * fix(opencode): select plugin default for execution host loader * fix(opencode): limit prompt prefill capability to verified release * feat(opencode): probe launch capabilities on the execution host * fix(opencode): select plugin loader for the launched host binary * fix(opencode): match WSL probe cwd and declared guest environment * fix(opencode): preserve launch environment deletion boundaries * wip(opencode): authorize native startup prompt intent at execution owner * fix(opencode): atomically replace status plugin entrypoints * fix(opencode): retain plugin permissions across restrictive umasks * test(opencode): resolve permission fixture from primary cwd * feat(opencode): install startup prompt plugin independently of status hooks * fix(opencode): wait for admitted startup intent and preserve failed-launch briefs * fix(opencode): unsubscribe hook settings during async host shutdown * fix(opencode): confine overlay manifest cleanup to owned directories Co-authored-by: Adnan Khan <adnank11427@gmail.com> * test(readiness): census recorded OpenCode composer boots * fix(opencode): reject redirected overlay parents before cleanup * fix(orcad): retain runtime cleanup when subscribing to hook settings * refactor(launch): extract OpenCode config and attachment authority * fix(opencode): retain host version selection across relay restarts * fix(opencode): pass run prompts as positional messages Preserve run flags and use the existing shell quoting and run-command detector to append the initial message after --, reusing an existing separator. TUI launches retain their version-selected prompt transport and draft behavior. Original run-order work: @coelho-doti (#13065, tracked in #17551). * fix(opencode): keep wrapped run tasks positional Recognize supported environment prefixes and PowerShell call operators without mistaking prompt arguments for executables. Keep environment and run separators separate, preserve the task text and exclude run commands from native submission. Source-parent: |
||
   
|
2fc517c1c6 |
feat(cli): add --unread/--read to orca worktree set
Pass validated read/unread flags through the existing workspace metadata update. Contributor attribution correction: the earlier Ctrl+M squash message placed its related link after the co-author footer, so GitHub did not recognize all contributors. Preserve that credit here for https://github.com/stablyai/orca/pull/24100 and its related contribution https://github.com/stablyai/orca/pull/24143 . Related CLI link validation retained from https://github.com/stablyai/orca/pull/20036 . Co-authored-by: Raz Shlomo <12373339+razshlomo@users.noreply.github.com> Co-authored-by: Neil <neil@stably.ai> Co-authored-by: Zhichang Yu <yuzhichang@gmail.com> Co-authored-by: Ahmed Nagy <ahmednagy25t@gmail.com> |
||
   
|
c1d403a47b |
feat(cli): link GitHub and GitLab items on worktree create and set
Expose and validate the existing workspace link fields, preserving omitted values and provider identity checks. Related: https://github.com/stablyai/orca/pull/22609 Co-authored-by: marco song <marco.song@mvlchain.io> Co-authored-by: SongMarco <20613630+SongMarco@users.noreply.github.com> Co-authored-by: Luca Critelli <lucacri@gmail.com> Co-authored-by: Neil <neil@stably.ai> |
||
|
Brennan Benson
|
a4606ccae3 |
fix(cli): orca file open no longer moves your view unless you pass --focus (#24244)
* docs(cli): file open/diff/open-changed say they switch the user's view and are for user requests only Refs #9944 * fix(cli): file open/diff/open-changed leave the user's view alone unless --focus `orca file open`, `file diff` and `file open-changed` always switched the desktop to the target worktree, selected the tab and revealed it in the sidebar. An agent skill that opens its answer pulled the user out of whatever they were typing in (#9944), and a phone opening a file moved the desktop too. The commands now add the tab in its worktree without changing anything on screen, including when that worktree is the one being viewed: the new tab is added to the tab bar but the active tab, tab type and focus stay put. In a worktree the user is not viewing, the tab becomes that worktree's selection so it is in front when they go there. `--focus` keeps today's behavior. files.open / files.openDiff take an optional `navigation` target (the existing RUNTIME_NAVIGATION_TARGETS vocabulary); the CLI sends 'all' for --focus, like `worktree create --activate`, and nothing otherwise. The renderer moves the host view only when the target reaches the host; a missing field (phones, older CLIs) leaves it still. Editor opens for a worktree other than the on-screen one no longer write the global activeFileId/activeTabType. Refs #9944 * test(cli): justify the window and runtime stubs in the file-open notification test * fix(cli): keep phone file opens switching the desktop; the CLI asks for 'caller' Phone opens send no `navigation` field, and the phone's diff-review "Open in session" relies on the desktop selecting the diff it opened. A missing field now keeps the original switch exactly; the CLI says what it wants instead: 'caller' (no host move) by default and 'all' for --focus. Older CLIs, which send nothing, keep switching as they always have. Refs #9944 * fix(cli): background file opens select the tab without counting as a visit A CLI open into a worktree the user is not viewing selected the new tab with the same activation a user click uses, which stamps lastFocusedAt and the group's recency list. The worktree jump palette sorts recent tabs by that time, so every agent `orca file open` into another worktree jumped to the top of the user's recent tabs. Editor opens now take a selection mode: 'focus' (default, unchanged), 'background' (select within its worktree without recording focus or recency) and 'none' (add only). createUnifiedTab and activateTab gain recordFocus:false for the background case. Also: tests for reopening an already-open file or diff without --focus, a comment that file opens move only the host window ('all' acts as 'host'), root help lines back under 100 columns, and an accurate remote test title. Refs #9944 * fix(tabs): a background-selected tab still joins its group's tab history recordFocus:false skipped both the focus-time stamp and the group's recentTabIds append while still making the tab the group's active tab. Ctrl+Tab looks the active tab up in that history, so after a background CLI open it did nothing (or went to the wrong tab) once the user switched to that worktree, and hydrate kept the broken history across a restart. Only the focus-time stamp is skipped now; the jump palette's recent rows sort by that alone, so the palette fix stands. Refs #9944 * fix(cli): file open/diff/open-changed --focus help says it brings the user to the file The three commands borrowed the shared --focus line written for terminal create ("Reveal the created terminal session in Orca"). They now use the per-command flag help table; terminal create's line is unchanged. Refs #9944 |
||
|
Neil
|
52a1e2875b |
feat(orchestration): accept Muse model and effort for supervised workers (#22383)
* feat(orchestration): accept Muse model and effort for supervised workers `worker-start --agent muse` already launched, but `--model` was refused because Muse had no session-option catalog. Add one that maps worker preferences to `muse --model <id>` and `--reasoning-effort <level>`; it seeds no models, so native-chat surfaces show no picker. opencode stays without `--model`: the opencode 2 TUI (now shipped as `opencode`) rejects the flag, so the refusal now tells callers to rely on the agent's own config. Help, skill guide, and docs list valid `--agent` ids and the agents that accept `--model`. Refs #19823 * test(mobile): repin session route closure for the Muse option catalog |
||
 Jinwoo Hong
|
4b1b7178ad |
fix(orchestration): scope @ group addresses to the sender's Run (#19783)
* fix(orchestration): scope @ group addresses to the sender's Run `@all`, `@idle`, and the agent-name groups (`@claude`, `@codex`, ...) resolved against every terminal on the host. A coordinator meaning "my three reviewers" reached 126 agents across every open project, twice in one day, and every unrelated agent burned a turn discarding mail that was never for it. Every group except `@worktree:<id>` now means the live Dispatches of the sender's own Run, each addressed as `dispatch:<id>` so delivery is durable even when the worker terminal is not attached yet. A sender bound to no Run is refused with `invalid_argument` naming `run:<id>` / `dispatch:<id>`; there is no host-wide fallback and the host's terminals are never enumerated for it. `@idle` and the agent-name groups filter within that set by the same terminal status and host-resolved identity as before. `ask --to @group` returns the same code and points at the owning Run mailbox. Federated Dispatches read relayed control mail rather than a local mailbox, so a Run-scoped fan-out skips them with a `recipient_unreachable` warning naming the direct `dispatch:<id>` address. Group addresses are resolved host-side, so no RPC or stream shape changes; an older CLI sending `@all` to a new host gets the Run-scoped meaning. Claude-Session: run-scoped-group-addresses * fix(orchestration): revalidate legacy takeover before the recipient verdict A legacy coordinator taken over while `listTerminals` was in flight reported `runtime_error` instead of `legacy_read_only`: Run scoping made "no live Dispatch in this Run" the first thing the group send could fail on, and that threw before the takeover check ran. Takeover is a precondition, not a commit-time detail — the sender must be told it is read-only whatever else is wrong with its recipient set. Revalidation moves to immediately after the only `await` in the path. Everything below it is synchronous, so the commit-time window it used to guard is unchanged; only the error paths now see it. The legacy partition test gave `term_current_worker` no Dispatch, so under Run scoping it is correctly not a recipient. It now holds a real current-contract Dispatch in the same adopted Run, which is what the test is named for: one `legacy_direct` and one `current_delivery` recipient in one fan-out. Claude-Session: run-scoped-group-addresses * fix(orchestration): address the Run a nested coordinator created, not its parent A nested coordinator is both a worker of its parent Run and the coordinator of the Run it created. `resolveMessageRun` answers with the parent, correctly, because that is where its own `worker_done` belongs — but audience is a different question. Scoping `@all` to that Run sent a nested coordinator's "shared context" to the siblings it was started beside instead of the workers it started, and reported success, so it never learned its sub-workers heard nothing. Before Run scoping the host-wide fan-out reached the sub-workers by accident; this turned an over-broad delivery into a wrong-audience one, the exact failure class the change exists to remove. Group audience now resolves off the Run the sender coordinates, falling back to its Dispatch's Run. A leaf worker coordinates nothing and is unaffected. This is a separate question from `routing.run`, not a second answer to the same one, so `resolveMessageRun` keeps its meaning for point-to-point mail. Also: when every live Dispatch in a Run is federated, the fan-out skipped them all and threw a bare `Error` that discarded the warnings naming those remote workers and how to address each one. The sender was told "no recipients" while three remote workers existed. That throw now carries a code and the skip explanations. Claude-Session: run-scoped-group-addresses * docs(orchestration): say that no group address reaches a coordinator A coordinator is not a Dispatch, so Run-scoped groups never include one. That follows from the rule, but nothing said it, and the old host-wide meaning did include the coordinator — a worker sending `@all` to raise a blocker would be heard by its siblings and by nobody who can act. The guide, the CLI note, and the docs page now say to use `run:<id>` for that, and that a worker which created its own Run addresses that Run's workers. Also restores the `@cursor` case dropped when the group tests moved: a Claude pane titled "Fix the text cursor blink" must not receive Cursor's mail. That hazard was recorded from real titles and `@droid` alone did not cover it. Claude-Session: run-scoped-group-addresses * fix(orchestration): preserve group audience and mailbox identity * fix(orchestration): validate group scope before dispatch routing * fix(orchestration): preserve pane identity and exclude coordinator dispatches |
||
|
Jinwoo Hong
|
06a607a1d7 |
feat(orchestration): make multi-agent workflows durable (#16904)
<!-- orca-pr-loc -->
<!-- Programmatic LoC summary. Do not edit by hand; rewritten on every commit. -->
| | Files | Added | Deleted | Net |
| :--- | ---: | ---: | ---: | ---: |
| Test | 225 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$21666 | $\color{#cf222e}{\Huge{\mathbf{−}}}$2820 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$18846 |
| Prod | 348 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$17107 | $\color{#cf222e}{\Huge{\mathbf{−}}}$4706 | $\color{#1a7f37}{\Huge{\mathbf{+}}}$12401 |
<!-- /orca-pr-loc -->
## ELI5
Orca now treats orchestration like a durable control plane instead of inferring success from terminal keystrokes. Agents can tell whether a prompt was accepted or a turn started, replay an ambiguous request without sending twice, and recover coordinator mail after a crash. Completed workers can be inspected, released, or retained, and their panes no longer auto-resume as if the work were still running.
## What changed
- **Run receipts** from `run-create/use/current/show/list` are the row without routing plumbing (`home_database`, `coordinator_pane_key`) and without the duplicate `binding` object.
- **`terminal send` receipts are honest and idempotent.** `input_accepted` and `turn_started` are the only stages; `--wait-submit` observes without resending; `--retry-request <uuid>` replays the exact request against the same process incarnation. A transport timeout keeps the retry ID; only a different runtime answering strips it. Value-less or non-UUID `--retry-request` is rejected on the CLI and the SSH shim.
- **Mailbox delivery is committed before wakeup.** Pointer writes are staged in the DB before any PTY byte, replayed once after restart, and never emit a naked Enter. The watermark that parks concurrent deliveries is released with the DB reservation. Restart rescans pointer-pending and `dispatch:` mailboxes.
- **Lifecycle is a guarded transition graph** (`lifecycle-transition.ts`) with a table-driven test over every caller edge. Task reopen/overturn stays in the public contract. A PTY exit during `worker-stop` is the stop succeeding, not a failure.
- **Worker lifecycle CLI:** `worker-start` (`--spec` creates Task + attempt in one call), `worker-show`, `worker-read` (provider transcript first, bounded terminal fallback with a typed reason, local/WSL/SSH), `worker-stop`, `worker-abandon`, `worker-release`, `worker-retain`, `worker-list` (rowid-fenced pagination, fleet liveness, `attention`, literal `nextAction`).
- **Release is an explicit ownership table** (`decideWorkerTerminalRelease`): only an `owned` resource can be settled, the archive is mandatory where reachable, and an owner whose process is proven exited can always get out of `retained` via `archive_status: unavailable`. User-taken-over, external, and transferred panes stay retained.
- **Settled-worker resume fence** (folds in #17651): a settled dispatch whose pane is still open is fenced at settlement, on stop/abandon/exit, and at startup; lifted on release, retain, takeover, and pane reuse.
- **Liveness is `live` / `unverifiable` / `exited` only**, from execution-host evidence. Fleet projection reads the evidence clock, not the relay delivery clock. A host-certified exit outranks the worker's settled state. `unverifiable` never authorizes stop, abandon, retry, or release, in code or in the guide.
- **Federation:** structured reads negotiate by `method_not_found` so every shipped host keeps transcript-first output; exited remote workers are closed before being reported closed; epoch fencing holds across peer restart, downgrade, and pairing rotation; no per-second forced capability probe.
- **Schema v35:** repairs databases stamped v34 by the pre-fix branch (mailbox_handle default, index predicates), drops the write-only `lifecycle_transition_receipts` ledger and five never-read v31 identity columns.
- **Schema v36:** `dispatch:<id>` mailboxes get a real consumer generation on `dispatch_contexts` and `remote_dispatch_attachments`, bumped and fenced in the same transaction on every re-attach (manual inject, worker-start, federated attach). A stale worker whose Dispatch moved to another process now gets `consumer_fenced` instead of silently acking the new worker's Delivery. Run mailboxes already worked this way.
- **Schema v37:** `dispatch_contexts` records its creator (`creator_handle`, `creator_pane_key`), so a coordinator's context-only self-dispatch is bookkeeping rather than a nesting parent; before this, one self-dispatch made every later `worker-start` from that coordinator fail the depth cap. Pre-v37 rows keep counting (fails closed).
- **Dispatch-mailbox ownership is checked, not inferred.** A `check` from a process whose pane no longer holds the Dispatch, or whose last Attempt was abandoned/failed and moved to another terminal, gets `consumer_fenced` instead of an empty inbox that reads as "no mail yet". `--peek`/`--all` stay readable. A paneless caller still gets `stable_pane_required` with the rebind recovery.
- **Liveness certification is stricter:** a `process_exited` stage whose termination reason is `unknown` (a stop that was issued but never observed) projects `unverifiable`, not `exited`. Federated `worker-show` carries the execution host's verdict and host kind instead of a local guess. A live, ready worker with nothing pending has `nextAction: none` rather than pointing at the `worker-show` that produced it.
- **Wire:** `workerShow` keeps `dispatch.task_id` next to `taskId` for shipped CLIs. `ask --json` uses the standard `{ok, result}` envelope like every sibling verb.
- **Migration start-version detection** treats the two v32 recovery columns as versioned. Before this, every shipped database stamped below 32 resolved to the v6 floor and replayed the whole chain (the v23 backfill synthesized 68 phantom retained workers on a real v30 profile). Verified on a copy of a real 62 MB v30 profile: starts at 30, no row delta, integrity ok, 11 ms.
- **Skill guide** rewritten as a ≤200-line kernel plus seven references, to the outcome-first standard (Result / Done / Safe failure first, conditions not case lists, one done bar, references loaded at the point of use). The canonical loop uses `worker-start --spec`, names `worker-list` for completion accounting, documents `--retry-request` / `request-show` / `--wait-submit`, and requires positive evidence before any stall action. The other seven guides get the same treatment in #18724, split out so this PR stays orchestration-only.
- **`rpc/methods/orchestration-*`** (126 flat files) regrouped into `orchestration/{worker,federation,messaging,runs,gates}/`.
## Why
User reports showed the same boundary failures: false `agent_prompt_stalled` causing duplicate sends (#15180), coordinators unable to trust screen scrapes, cold-parked terminals receiving a pointer without the submit, settled workers accumulating as live tabs and auto-resuming after restart, and no way to tell a stalled worker from a working one.
## Linked issues
Fixes #15180. Fixes #17935 (orchestration skill description is 866 characters; a guard now caps every bundled skill at 1,024). Supersedes #17651 (fence folded in). Advances #16660, #16522, #14907, #13047.
## Review record
This PR was reviewed adversarially after revival: eight independent lenses (lifecycle, mailbox, send, worker, federation, transcript, complexity, live ergonomics), each required to prove findings with a failing test. That produced 16 proven blockers, all fixed with red-then-green regression tests, followed by two re-review rounds and a third fix wave that caught 3 regressions introduced by the fixes and 7 fixes that missed their target; all closed. A final pass (five lenses incl. a live built-runtime smoke, then a re-review of the fix wave) found and fixed seven more, chiefly the stale-worker mailbox steal, the self-dispatch depth wedge, and the unproven-exit certification. Three independent Codex (gpt-6-astra) passes followed: the first found nothing new, the second found and fixed 3 defects (task-status reachability, WSL-local host classification, peer-capability epoch), the third found and fixed 6 (production PTY controller never installed settled writes, ambiguous in-flight pointer failures allowed duplicate replay, SSH/relay deadlines cut off a valid `--wait-submit`, stop-vs-exit race during inspection, and two release-recovery paths for vanished or exited terminals). The full record (findings, proof tests, triage, declines with reasons) is archived outside the repo.
**Rework after the live smoke.** A first live cross-host run on the shipped adhoc build (this Mac, a paired Windows host on the same build, a paired Mac on 1.4.195, and an SSH host) found a P1: a running local worker read `unverifiable`/`missing_status` because the fleet snapshot rows lacked the terminal handle the matcher keyed on. A 59-row failure table over every bug fixed during review showed the same two classes recurring: a fact dropped in transit through optional fields, and two authorities for one fact. Two blind designs (Opus, Codex) converged on the same mechanisms, and the scoped tranches landed here with red-then-green seam tests from the real producer to the real consumer, faults injected only at the transport or hook-ingest boundary:
- **Settlement (data-loss class):** one three-valued `WriteSettlement` (`accepted | refused{reason} | unverifiable{reason, bytesHandedToTransport}`) from the SSH multiplexer through daemon client, providers, controller, to pointer staging. No boolean, no rejection-as-third-state. The two silent degrades that fabricated a handoff are deleted; a provider that cannot settle refuses before any effect. Pointer text and Enter share the contract; a partial flush is `unverifiable`, never `refused`.
- **Evidence identity (false-liveness class):** fleet agent-status evidence is a tagged union (`binding: worker | pane | unresolved{reason}`, `clock: observed | delivery`) minted once at ingest, so a hook row captured on one process incarnation can never bind to a later dispatch on the same pane. The matcher's `!worker.paneKey ||` defaults are gone. One host-scope parser replaces two.
- **Small pre-merge items:** `capability_unsupported` from an old peer is no longer relabelled `host_unavailable`; a producer census test asserts every agent-status consumer path projects a pane-only hook row as `live`.
Two ergonomics defects the second live run surfaced on a real database are fixed here too: a pre-v3 dispatch already marked `completed` projected as `outcome_unknown` / `requiresAction: true` forever (three copies of the outcome ladder disagreed on legacy rows; now one resolver, legacy `completed` reads `succeeded` with nothing to act on, legacy `failed` stays actionable on the failure), and an unscoped `worker-list` enumerated the entire database (now defaults to the Run bound to the calling terminal, `--run` overrides, and the receipt's additive `scope` field says which).
A third live round on the shipped adhoc build of `b082443e1f` (same four hosts) plus an unscripted run in the user's own prompt style (a plain Claude Code shell, `/orchestration`, three workers, zero errors, bound-Run default confirmed) found two more branch defects, fixed with red-then-green tests: a worker freshly started on a paired server projected `unverifiable`/`host_indeterminate` with `requiresAction` for ~3 minutes, including after its own `worker_done`, because the host's federation observation returned `missing_liveness_verdict` for any PTY the liveness register had not yet swept (the host now reads a connected pane it owns locally as `live`; disconnected or SSH-scoped panes stay `unverifiable`); and six pre-v3 completed rows still carried an `input` category because settling through the task-status path or `failDispatch` never closed the Dispatch's pending question threads (both paths close them now, and schema v38 closes threads already pending on settled rows). The guide's `worker-start` examples now show `--model sonnet`, since an omitted model inherits the launcher's default.
A Codex adversarial pass on the tranche diff found one real design hole (identity minted at read time instead of ingest, now closed) and two daemon settlement paths that threw instead of settling (fixed). Two `@ts-nocheck` runtime mixins on these paths were extracted into checked modules; the repo-wide `@ts-nocheck` count is unchanged at 171.
Deletions during review: ~1,900 lines (write-only ledger, unread columns, dead v1 archive path, test harnesses shipped in prod, duplicated liveness and state-machine copies, self-capability checks that were compile-time true).
## Testing
- `pnpm typecheck:tsc:node|cli|web` clean
- `pnpm run check:code-quality:changed` 0 findings; `check:react-doctor:changed` 0
- `pnpm verify:bundled-skill-guides`, `verify:skill-bundle-manifest`
- full `pnpm test` on the integrated head: 72,332 pass / 292 skipped; the only failures were three non-PR files (two zsh live-shell suites hit a node-pty spawn-helper ENOENT while a concurrent native rebuild ran, 44/44 in isolation; `release-checkout.unit.test.ts` is a known 30 s load timeout that passes in isolation on `origin/main` too).
- CI on
|
||
|
Neil
|
51eed5a1bc |
feat(cli): report SSH host platforms (#18896)
* feat(cli): report SSH host platforms * feat(cli): include SSH connection status * fix(cli): preserve unknown SSH connection state |
||
|
Jinwoo Hong
|
573537ecd4 |
feat(cli): make terminal close the canonical workspace teardown (#18073)
* fix(runtime): recover stale session owners and await retirement * fix(runtime): preserve session hydration and smoke compatibility * test(runtime): cover empty and unindexed session owners * feat(cli): make terminal close the canonical workspace teardown * fix(preload): align ssh termination result type * test(runtime): assert folder hydration owner * fix(runtime): fence legacy terminal stop by worktree host * fix(preload): reconcile ssh result import with main * fix(runtime): keep same-id sibling hosts out of workspace close The stale-owner fallback in the session controller re-routed any worktree whose catalog partition had no tabs to whichever other partition held tabs. Only `runtime:` environment ids rotate across relay restarts; `repoId::path` legitimately repeats across hosts, so an SSH workspace close could retire the local copy's tabs and resume records, or flip owners mid-close and strand the SSH PTY. Restrict the fallback to runtime hosts, and pin the session partition once per workspace close so record clearing targets the partition that owned the tabs. * test(runtime): give the cross-host close fixture a real resume record * fix(preload): take main's ssh-bridge import order so the merge stays duplicate-free |
||
|
Neil
|
e5a1e79e8e |
docs(linux): say which package to install and how updates arrive (#18123)
* docs(linux): say which package to install and how updates arrive Closes #5188. Closes #10987. The install guide's entire Linux section was "AppImage and `.deb` builds are available. See the Releases page for details." It named two of the three published packages, gave no basis for choosing between them, and said nothing about updating -- which is the one thing that actually differs between them. Separately, nothing human-facing said the Linux CLI is `orca-ide`; only skills/orca-cli/SKILL.md carried it, which agents read and humans do not. Install page now picks the package by update behaviour: the AppImage self-updates, deb/rpm report the new version and hand over the install command, and a repackaged build is not offered a download it cannot apply. Records that Orca never escalates privileges for the package install, and points at #18086 for the signed repo as planned, not shipped. Adds .rpm to the download list. Release CI builds it (release-cut.yml: `--linux AppImage deb rpm`) and verify-release-required-assets.mjs requires the artifact, so omitting it was just wrong. The CLI command name is now stated where humans hit it -- the CLI reference and overview -- with the GNOME Orca collision as the reason, plus the two places bare `orca` does work: inside Orca-managed terminals (PTY PATH shim) and on a packaged `orca serve` host (the ~/.local/bin dispatcher). The headless guide gains the same note, which is what makes its `orca skills install` lines correct rather than a typo. * docs(linux): fix install ordering, CLI verification, and serve bootstrap Readiness review found ten defects. Two would have had a reader run the wrong program, and one would have had them install a .deb over a live app. Install ordering was reversed. The page said "run it, then quit and reopen Orca"; the ref this is gated to land with says the opposite in four places (linux-package-downloaded-status.ts LINUX_PACKAGE_MANUAL_INSTALL_MESSAGE, "Quit Orca before running the system package install command", plus the recovery card's title, summary and explainer). That wording came from main's older run-then-quit card, which the stack deliberately reversed when it retitled the card to "Manual Install Required". Now: quit first. CLI verification put the Linux caveat *below* `command -v orca`. That check succeeds on any GNOME desktop and resolves to the screen reader, so the reader got a confident hit from the page's own verification step and then invoked the wrong program. Caveat moved above, and the block now spells `orca-ide` literally instead of asking the reader to substitute. The serve bootstrap was circular: the bare-`orca` dispatcher is written *during* serve startup (main-process-runtime-launch.ts), so it can never be the command that starts serve. First launch is `orca-ide serve`. Fixed here and in the two pages this links to. Accuracy: the install command now matches what the code emits -- absolute paths resolved from the trusted directories and a POSIX-single-quoted package path, as pinned by linux-package-install-command.test.ts -- and names the manager fallbacks (dpkg; zypper/dnf/yum/rpm) rather than presenting apt as the only form. The pending path honours XDG_CACHE_HOME. rpm arch tokens are x86_64 and aarch64, not deb's amd64/arm64. arm64 AppImage is linked. Dropped the container example: isExternallyManagedLinuxInstall() needs a root marker AND no trusted package manager, and a Debian-based container has apt, so it is not flagged. |
||
|
Jinjing
|
45c4823109 |
Format documentation with consistent line wrapping and table alignment (#17765)
Standardize MDX files across docs with: - Remove trailing semicolons from import statements - Wrap long lines and multi-line component props for readability - Align Markdown table column separators - Normalize text and JSX formatting for consistency |
||
|
Neil
|
6aba202d5e |
feat(docs): publish OSS docs with stable releases
Publish the standalone docs site under docs/site and deploy it on stable desktop releases. |
