The workspace link means pnpm never creates a .pnpm/@orca+windows-registry@*
entry, so all four native-cache blocks globbed a path that cannot exist and
the addon was recompiled on every Windows job.
Also hardens the addon itself: RegEnumValueW reports a byte count and the
registry does not enforce whole WCHARs for string types, so an odd count let
Napi's auto-length scan run past the value; and a value named __proto__ would
reassign the result object's prototype instead of becoming an entry.
* refactor(windows): vendor the registry addon as @orca/windows-registry
windows-native-registry@3.2.2 was last published in 2023 by a single
maintainer. Orca called two of its exports, both read-only, so the whole
dependency is replaced by a local N-API addon under native/.
The vendored addon is read-only by construction: setValue, createKey and
deleteKey are gone, so RegDeleteTreeW no longer ships in the app. Two
upstream defects are also fixed rather than carried over — the name/data
scratch buffers were file-scope statics that concurrent reads would
scribble over, and createKey/deleteKey called .c_str() on a temporary.
Build wiring keeps the existing shape: still an optionalDependency gated
to win32, still excluded from pnpm's allowBuilds so only Orca's own
Windows rebuild runs node-gyp for it, still copied into the packaged
resources. The CI native caches now key on the vendored sources so an
addon.cc edit cannot restore a stale .node.
* test(windows): check the vendored registry addon against reg.exe
The addon is vendored source, so no upstream release proves it still
decodes values the way Orca's PATH readers expect. reg.exe is the only
independent oracle on the box.
* ci(windows): register the registry addon test on the Windows runner
A Windows-gated file self-skips on ubuntu, so without both registrations
it reports success while running on no machine at all.
* fix(build): link the registry addon as a workspace package, not file:
As a `file:` dependency pnpm re-resolved and re-linked the package on
every install, including `--frozen-lockfile` (measured: "added 1" on a
repeat no-op install). That virtual-store churn ran concurrently with
node-gyp reading the same tree and cost @vscode/windows-process-tree its
binding.gyp mid-rebuild, failing package (windows) whenever the native
cache hit and only that module needed building. The linux packaging job
hit the same race from the other side, as a pnpm staging move failure.
A workspace link resolves once and leaves the store alone; repeat
installs are now 55ms no-ops. native/windows-registry is listed
explicitly so `packages:` still does not auto-discover mobile/.
* fix(build): stop tracking node-gyp output for the vendored addon
The build/ tree is generated per host and ABI; the committed copy was
macOS-specific gyp scaffolding from a local build and would have shipped
stale Makefiles to every checkout.
* chore: ignore the vendored addon's node-gyp bin output too
node-gyp also emits bin/<platform>-<abi>/ beside build/; both are per-host
generated output that must never be committed.
* fix(computer-use): run the Windows runtime as one persistent helper
Microsoft Defender for Endpoint raised multi-stage Execution + Collection
incidents against Orca on Windows ("Screenshots were taken unexpectedly on
this device... Screen capture code was found in a script launched by
powershell.exe", factor "Executes suspicious MSIL code"). The desktop script
provider spawned a fresh powershell.exe per operation, so a single computer-use
session produced a burst of short-lived PIDs and re-emitted runtime.ps1's
inline Add-Type P/Invoke assembly on every click.
runtime.ps1 gains a -Serve mode that loads its assemblies once and then reads
NDJSON requests from stdin, and a new DesktopScriptRuntimeHost owns one
long-lived child: lazy spawn, strict serialization, a 30s per-request timeout,
restart on crash, a 120s idle shutdown, and dispose() on provider teardown. The
one-shot -OperationPath path stays as the fallback, and Linux keeps its python3
bridge unchanged.
Both Windows spawn sites now use -ExecutionPolicy RemoteSigned instead of
Bypass, falling back once to Bypass (and logging) when a Restricted host
refuses the unsigned script.
* fix(computer-use): recover the runtime host instead of latching it off
Review follow-up on the persistent Windows computer-use helper.
A helper that died before producing a line set an unavailable flag nothing ever
cleared, and the client then dropped the host for the life of the session. One
transient bad spawn — a Defender scan, a locked CSC temp directory — silently
restored the per-click powershell.exe burst and per-operation MSIL emission this
work exists to remove, with computer use still working so nothing looked wrong.
Start failures are now retried, then cool down for 60s, then re-probed; the
client keeps the host so it can come back. Repeated post-answer crashes cool
down too, and a single reply no longer clears the failure count.
The one-shot bridge decided its execution-policy retry from a message that fell
back to stdout, so a window title containing "SecurityError" could replay a
non-idempotent operation — a double click, keystroke or paste — and stick the
session on Bypass. The retry now requires empty stdout and a matching stderr.
Serve-mode replies carry an echoed request id. Without one a single stray stdout
line would make every later response answer the previous request, acting on
stale element indexes with no error raised; a mismatch now kills the child.
Non-JSON noise is ignored rather than counted as the helper having answered.
Also: warnings reach the main process over the sidecar's IPC channel rather than
its piped, unread stdio; the child is watched on close rather than exit; dispose
latches so a queued request cannot respawn during teardown; and the host is
split into a serve channel and an availability policy to stay under max-lines.
* fix(computer-use): prove a helper never started before replaying its request
The retry that replaced the permanent-latch bug could deliver unrequested
input. send() re-sent the same request whenever the helper died without
replying, but "no reply came back" is not "the operation did not run":
runtime.ps1 synthesizes the click and only then builds the snapshot, which
allocates a full-window bitmap and walks the UIA tree — a native GDI+/UIA fault
there is uncatchable, and leaves the click already delivered. A deterministic
fault meant three clicks from the host plus a fourth from the one-shot bridge,
surfaced as a single failed operation.
-Serve now writes one {"ready":true} line after its Add-Type work and before
its first read, so "never started" is a fact rather than an inference. A request
is replayed only when the helper died before announcing. A runtime.ps1 that
predates the announcement — reachable through the provider path override — is
covered by an observation-tool allowlist until a ready line proves otherwise.
Host-detected aborts (timeout, desynchronised reply, oversized line) suppress
the exit handler, so they were bypassing failure accounting entirely and a
helper failing that way was respawned once per operation forever. They now
count and are logged.
Also stop charging twice for one outage: entering the cooldown resets the
failure count, so the first death after recovery no longer re-enters a full
cooldown and an interleaved workload cannot be stranded on the one-shot bridge.
* fix(computer-use): ignore a stdin write callback from a torn-down helper
stop() destroys stdin, so a write still queued at teardown calls back with
ERR_STREAM_DESTROYED. The callback carried no channel or request identity and
write() had no closed guard, so it ran abortChannel a second time: stopChannel
no-opped but recordFailure and the warning did not, charging two failures for
one operation and reaching the 3-strike cooldown at half the intended rate.
That feeds the same accounting that keeps a persistently broken helper from
respawning once per operation.
The same root also allowed a late callback landing after a replacement channel
existed to stop that channel and reject a different request with the previous
one's error. Node fires the destroyed-stream callback on the next tick, well
before a new request arrives, so the double-count is the reachable effect;
binding the callback closes both.
write() now drops payloads and error reports once closed, and the host ignores
any report whose channel or request id is no longer current.
* test(computer-use): pin each stale-write guard independently
The channel's closed guard and the host's request-identity check are redundant
by design, and the existing tests only failed when both were absent. Someone
deleting one, believing the other was the covered one, would have got a green
suite and a live regression — the same shape as a test that passes without the
fix it was written for.
Each is now pinned on its own. The channel's half is tested against the channel
directly: after stop() it takes no writes and reports no error from one already
queued, which the host cannot observe because it drops the channel at the same
moment. The host's half is pinned by the case the channel cannot see — a live
channel whose request was already answered, where backpressure delivers a write
callback for a request that is no longer pending.
Removing either guard alone now fails a test. Both carry a comment saying they
are deliberately redundant and separately pinned, so the next reader does not
have to rediscover this from the diff.
* ci(windows): run the computer-use runtime host suite in CI
The win32 suite only self-skips off Windows, so it passed vacuously in
every lane. Register it the way the cmd-shim suite is registered.
* fix(computer-use): time the runtime host cooldown on a monotonic clock
The start-failure cooldown was a wall-clock deadline, so a backwards step —
an NTP correction, a VM snapshot restore, a user changing the clock — left
`remainingCooldown()` returning the cooldown plus the whole step. A one-hour
step measured 3,660,000ms, and ten real minutes later still 3,060,000ms.
Nothing shortens it from there. Only `recordSuccess()` clears the cooldown on
a non-dispose path, and no request can reach a helper to succeed while it
holds, so every `send()` throws `runtime_host_unavailable` first. The host is
built with no `now` override and its lifecycle is a module-level singleton
that shuts down at process exit, so the latch held for the sidecar's life —
computer use kept working via the one-shot bridge while the per-click
powershell.exe burst this host exists to remove came back silently.
Store the instant the cooldown began and compare elapsed monotonic time,
following the two fixes in #17884. The field is `number | null` rather than
sentinel 0 because `performance.now()` legitimately returns 0.
Both new tests leave `now` unset, because the bug was in the default the host
picks and a test that injects a clock cannot see it.
* fix(computer-use): give a queued request its own deadline
The 30s request timeout was armed only in `sendOnce`, once a request reached
a helper. A request behind N timing-out ones therefore waited roughly N times
that with no deadline of its own: bounded, but the caller sees an `await` that
looks hung for minutes and gets no error to act on.
Move the serialization tail into its own class and arm a deadline at enqueue
time. Only the wait is bounded — a request that reaches a helper still gets
its full execution budget, so nothing that used to succeed now fails. An
expired request is dropped rather than sent late: the caller has already been
told it failed, and a click delivered after that is worse than no click.
The tail keeps its never-rejecting shape and chains on the turn rather than on
the raced promise, so a caller giving up early cannot release the next request
while its predecessor is still in flight.
* fix(computer-use): stop reading a locked file as an execution policy block
`UnauthorizedAccess` is the FullyQualifiedErrorId PowerShell reports for a
policy block, and it is also a strict prefix of `UnauthorizedAccessException`,
which .NET raises for any ordinary locked or ACL-denied file. The predicate
matched the token unanchored, so an AV scan holding runtime.ps1 or a locked
CSC temp directory was read as a policy block.
Two consequences, both bad. `escalateExecutionPolicy()` has no path back, so
one false match spent the rest of the session on `-ExecutionPolicy Bypass` —
the exact command line token this stack exists to stop emitting. And on the
one-shot path `isPolicyBlockedStart` re-runs the operation: one-shot mode
writes stdout only after the operation returns, so a crash partway through an
action is indistinguishable from a helper that never started, and the click
lands twice.
Measured on Windows against all three records, which the test carries verbatim
as fixtures:
policy/Restricted FullyQualifiedErrorId: UnauthorizedAccess
policy/RemoteSigned FullyQualifiedErrorId: UnauthorizedAccess
genuine access denied FullyQualifiedErrorId: UnauthorizedAccessException
`\b` is the whole discriminator: between `s` and `E` both sides are word
characters, so no boundary exists there and the exception cannot match.
Dropped two alternatives that measurement showed were wrong. `PSSecurityException`
never appears — the record surfaces through a native-command wrapper and reports
`ParentContainsErrorRecordException`. The prose is wrong three times over: it
differs by policy, it is localized, and PowerShell hard-wraps it mid-sentence.
Anchoring on the `FullyQualifiedErrorId:`/`CategoryInfo:` labels would be more
precise again, but those labels are localized where the values are not, so it
would lose a real block on a non-English host and strand it with no fallback.
Matching the values with word boundaries keeps both directions; a fixture with
translated labels pins it.
The escalation stays sticky. With the predicate correct, it only fires on a
machine that really does block, where re-probing the preferred policy would buy
a guaranteed failed spawn per operation.
* fix(computer-use): route a malformed request back to the request that caused it
`ConvertFrom-Json` throws before `$requestId` is read, so the serve loop
answered an unparseable request with an untagged error. On the client that is
not an error at all: `deliver()` sees no matching id, calls `abortChannel`,
kills the helper and charges a failure — and the helper's own message is
discarded. A parse failure was reported as a stream desync with no trace of
the real cause, and three of them walked into the 60s cooldown behind three
misleading "did not match" messages.
Recover the id from the raw line when the parse fails. No wire change: the
response shape is untouched and `BridgeResponse.requestId` already documents
this echo. It is the same shape the helper already returns for `not_a_tool`,
where the id survives because it is read before the operation runs. Both
mixed pairings degrade safely — a new script with an old client resolves the
error normally, and an old script with a new client still aborts, but now
reports what the helper said.
When the line is mangled past recovering an id, the desync abort is the honest
outcome, so keep it and carry the helper's text into it rather than replacing
it. A line the helper could not tag is usually the only account of the cause.
Proven against the real `runtime.ps1 -Serve`: the host can only write
well-formed JSON, so the parse-failure branch is unreachable through it and
the test drives the channel directly.
* fix(computer-use): keep the Bypass escalation only when Bypass actually works
AppLocker and WDAC constrained language mode raise PSSecurityException under
the same SecurityError category a real execution-policy block uses, so the
predicate matches them - correctly, on the evidence available. But those block
the script at parse time, which `-ExecutionPolicy Bypass` cannot lift. The
escalation was sticky unconditionally, so on a WDAC host we misdiagnosed,
retried, failed again, and then latched: every later command line carried the
most heavily weighted MDE token there is, on exactly the hardened, monitored
enterprise machine that is watching for it.
Treat the escalation as the diagnosis it is. A fallback that cannot start a
helper either disproves it - the policy was not what stopped the first attempt
- so revert to RemoteSigned instead of latching. When Bypass does start a
helper the diagnosis is confirmed and it stays sticky exactly as before, so a
genuinely Restricted machine still never pays a re-probe per operation.
The revert lands inside the outage rather than only at its end, so a
misdiagnosis costs one Bypass command line instead of one per attempt, and an
escalation that never proved itself does not outlive the cooldown that ends
the outage. Deliberately not a permanent "fallback is useless" flag: a Bypass
attempt that failed for a transient reason would then disable the fallback for
the session, which is the same latch in the other direction.
Only `runtime_host_unavailable` proves no helper started, so only that reverts;
a helper that started and then died proves Bypass works. That also makes the
policy branch reachable on a final attempt for the first time, so it now
rejects as unavailable rather than a generic error - that code is what routes
the operation to the one-shot bridge, which carries its own policy fallback,
and without it an all-blocked host would fail operations outright instead of
degrading. The pre-existing "reports itself unavailable when Bypass is also
refused" test pins that.
---------
Co-authored-by: Orca Worker <orca-worker@localhost>
* fix(terminal): preserve Option-composed ASCII input
* fix(terminal): preserve Option keyboard protocol semantics
* fix(terminal): complete Option keyboard event encoding
* fix(terminal): harden Option input encoding
* fix(terminal): close keyboard protocol fallback gaps
* test(terminal): prove Option-composed ASCII reaches the pty end to end
The Option-compose fix had unit coverage only. This drives a live Electron
pane whose kitty flags are armed by the application's own CSI > 1 u and
asserts the bytes at the pty boundary: composed `@` and Shift-layer `\`
arrive as text, configured Option-as-Alt still reports the layout-resolved
chord, and a non-ASCII glyph still reaches the app as its alt hotkey.
Restoring the pre-fix policy fails exactly the two composed-text scenarios.
Also records the ASCII rule's rationale where the rule lives, not only in a
test comment.
* refactor(terminal): drop the unread Option layers from the layout snapshot
The native helper computed an Option and Option+Shift character for every
key, shipped both over IPC, validated them in the parser and cached them in
the renderer — but no production caller ever asked for them. Only the base
and Shift layers are read, and Shift is the one the web layout map cannot
supply, which is why the helper exists at all.
Removing them halves the helper's UCKeyTranslate work per key and drops the
option parameter that six signatures were threading through for nobody.
* feat(computer-use): support macOS middle click and gate the AX click path
`--mouse-button middle` already validated end-to-end through the CLI, the
zod schema, and the provider validator, and both the Windows and Linux
providers honored it. Only the macOS provider rejected it outright with
"middle-click is not yet supported", so the flag was a dead end on the one
platform that has no fallback.
Two changes:
- Add `.middle` to the macOS button mapping. macOS has no dedicated middle
event family, so it rides `otherMouseDown`/`otherMouseUp` with the button
number carried by `mouseButton: .center`; that constructor argument is
honored for exactly the `otherMouse*` types, so no extra field write is
needed.
- Validate the requested button before the accessibility fast path, and skip
that path for buttons it cannot express. Previously the raw string was read
unvalidated, and `performClickAction` only special-cased `right`, so
`click --mouse-button middle --element-index N` (no modifiers, count 1) fell
through to `AXPress` — a left click — and reported success with
`path: "accessibility"`. Any unrecognized button string did the same. This
matches guards the Windows and Linux providers already had.
The button enum moves into `OrcaComputerUseMacOSCore` so it is unit-testable;
`main.swift` keeps only the CoreGraphics mapping.
Also documents `--mouse-button` in the computer-use skill guide, which never
mentioned the flag, so agents on Windows and Linux had no way to discover it.
* test(computer-use): cover macOS middle click in the real-desktop e2e suite
* test(computer-use): prove macOS middle-click delivery
Mouse events posted with CGEventPostToPid reach the target app with no
window association, so AppKit never routes the press to a view: hover
states fire but the control is never activated, and the mouseUp is
dropped outright when posted back-to-back. Post click events to the HID
event tap instead (as keyboard synthesis already does), pace them, and
stamp mouseEventClickState so multi-clicks register.
Synthetic clicks now also report verification unverified/synthetic_input
from the helper itself, matching the other synthetic actions.
* fix(windows): stop the Orca CLI dying on a duplicated PATH/Path environment
The packaged Windows `orca.exe` launcher read
`ProcessStartInfo.EnvironmentVariables`, whose lazy getter copies the
case-sensitive process block into a case-insensitive dictionary via `.Add`.
An inherited block carrying both `PATH` and `Path` threw
`ArgumentException: Item has already been added. Key in dictionary: 'PATH'`,
so every `orca` invocation exited 1 before Electron started
(native/windows-cli-launcher/OrcaCliLauncher.cs:46, printed at :67).
The launcher now mutates its own environment with
`Environment.SetEnvironmentVariable` and never touches either
`ProcessStartInfo` env property, so `CreateProcess` passes a NULL environment
block and the child inherits the live one verbatim.
Orca was also minting the duplicate itself. `applyTerminalAttributionEnv`
read `baseEnv.PATH` and unconditionally wrote `baseEnv.PATH`, so a Windows
PTY that inherited `Path` got a second spelling; which one the child resolved
was non-deterministic. `createLaunchEnv` did the same and, because its read
always missed on Windows, shipped Agent Teams terminals a `PATH` containing
only the tmux shim dir.
`resolvePathEnvKey` (extracted from the existing precedent in
windows-environment-path.ts) now drives every PATH read and write in the PTY
env pipeline, and attribution collapses Windows onto the single OS-resolved
spelling. Off Windows the resolver always returns `PATH`, so POSIX behavior
is unchanged and a case-sensitive POSIX `Path` variable is never touched.
Closes#12046
* test(windows): track the launcher's own-environment marker
The #12046 fix moved ORCA_WINDOWS_PACKAGED_CLI_LAUNCHER and ORCA_CLI_COMMAND
off ProcessStartInfo.EnvironmentVariables, but this asset test still pinned the
old dictionary writes and failed.
Co-authored-by: Orca <help@stably.ai>
* fix(windows): follow the host block's PATH spelling on sparse daemon env patches
Resolving a path-less Windows env to `Path` handed the daemon's own
`{...process.env, ...opts.env}` merge both spellings when the host block spelt
`PATH`. Fall back to the host block's own key, and collapse again inside the
daemon since that merge happens after attribution.
Co-authored-by: Orca <help@stably.ai>
* fix(windows): resolve the live PATH spelling by block order, not casing
Win32 resolves a duplicated variable by taking the first case-insensitive
match in the block, so `resolvePathEnvKey`'s hardcoded `Path`-first
preference targeted the shadowed spelling on the reporter's own
`["PATH","Path"]` block. Drop the attribution-side collapse with it: it
deleted the other spelling's value, and deleting the live key promotes
the shadowed one, so an env that stripped down to empty lost both.
* chore: drop unrelated merge formatting
---------
Co-authored-by: Orca <help@stably.ai>
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
* perf(computer): add mac helper owner-loss benchmark
Measure the release helper's resident memory before and after its owner-session deadline. Record exact revisions, per-trial RSS, retained state, and clean-exit latency so lifecycle reclamation is reproducible.
* fix(computer): reap mac helper after client loss
Bind the detached macOS helper lifetime to authenticated socket ownership. Reap the helper after its final authenticated client disconnects, and add a startup deadline for sessions that never authenticate.
* test(computer): harden owner benchmark cleanup
* test(computer): make owner benchmark cleanup failure-safe
* test(computer): close remaining owner cleanup races
* feat(onboarding): state-aware macOS notification permission step
The Set up notifications step showed a one-size-fits-all 'Open Mac
Settings' button that simultaneously fired the macOS permission prompt
and opened System Settings — two competing system UIs, with System
Settings unnecessary for the common fresh-install case.
Electron exposes no API to read macOS notification authorization, but
scheduling outcomes do reveal it: a silent probe notification's 'show'
event means permission is granted, 'failed' means delivery is blocked.
A new notifications:probeDelivery IPC runs that probe (cached via
passive delivery evidence and a persisted confirmation flag), and the
onboarding card now renders the real state:
- fresh install: the probe itself pops the native Allow dialog the
moment the step opens; the card flips to 'Notifications are enabled'
automatically when the user clicks Allow (silent 2.5s re-probes)
- blocked: amber card with an Open System Settings deep-link, which
also self-heals once the user flips the toggle
- granted: green confirmation card
The test-notification button now feeds the same card instead of the
ambiguous 'if no banner appeared…' toast during onboarding.
Co-authored-by: Orca <help@stably.ai>
* fix: don't log expected probe rejections while polling for permission
Co-authored-by: Orca <help@stably.ai>
* fix: amber warning styling + single stable dev bundle id for notifications
- Blocked card now uses the app's shipped amber idiom (tinted surface with
amber title/body) instead of white-on-amber-wash, which read muddy in
dark mode; macOS permission card split into its own module to stay under
the max-lines budget.
- Dev instances previously minted a unique macOS bundle id per
branch x Electron version, registering a new Notification Settings entry
every time ('Orca: <branch>' rows piling up forever) and pointing the
settings deep-link at ids System Settings can't resolve. All dev
instances now share com.stablyai.orca.dev: one Notification Center
entry, one permission grant covering every dev build.
Co-authored-by: Orca <help@stably.ai>
* fix: tighten macOS permission card copy
Body copy was one long sentence; now a single short instruction with
'Updates automatically.' as a separate dimmer line. Also repairs locale
catalog parity for keys introduced by commits rebased into this branch.
Co-authored-by: Orca <help@stably.ai>
* fix: drop 'Updates automatically.' line; ad-hoc sign dev app copies
The extra line read as confusing filler — the cards now carry one short
instruction each.
Dev Electron copies had broken code signatures (the Info.plist identity
edits invalidate the ad-hoc seal), which macOS punishes by refusing
Notification Center registration outright: every dev notification failed
with UNErrorDomain error 1, the app never appeared in System Settings >
Notifications, and the settings deep-link had nothing to land on. The dev
runner now ad-hoc re-signs the copied bundle after the plist edits
(bundleLayoutVersion bumped so stale unsigned copies are recreated).
Verified end-to-end: runner-built copy passes codesign --verify --deep,
probe delivery returns delivered, the onboarding card flips green in dev,
and the deep link opens the dev app's own notifications pane.
Co-authored-by: Orca <help@stably.ai>
* fix: drop confusing copy line; session-only permission evidence
Removes the 'Updates automatically.' line from both permission cards.
Also drops the persisted notificationDeliveryConfirmed flag: OS-level
permission changes between sessions, and a stale positive rendered a
false green card. Delivery evidence is now session-scoped only.
Documented detection ceiling (verified empirically on macOS 26): while
the permission dialog is unanswered — and when notifications are toggled
off in System Settings after being authorized — macOS accepts requests
and silently swallows them, with no public API (Notification Center
delivered-history and legacy ncprefs both included) able to distinguish
that from real delivery. 'failed' remains definitive for unsigned builds
and dialog-level denials.
Co-authored-by: Orca <help@stably.ai>
* feat: real macOS notification permission readout via native helper
Electron has no API for UNUserNotificationCenter authorization, and every
observable fallback lies: scheduling succeeds (and getHistory lists the
notification) even while macOS silently swallows display because the
permission dialog is unanswered or notifications were toggled off in
System Settings. The onboarding card therefore showed 'enabled' after the
user disabled notifications.
Adds native/notification-status-macos: a tiny Swift binary that prints
the app's real authorization status. It runs from inside the app bundle
(NSBundle resolves the bundle by walking up from the executable) and
embeds the app's CFBundleIdentifier in a __TEXT,__info_plist section so
every codesign --force pass — electron-builder's signing or the dev
runner's ad-hoc deep sign — derives the identifier macOS keys
notification records to. Spawning it from the app returns authorized /
denied / not-determined exactly matching System Settings.
notifications:probeDelivery now prefers this readout (authoritative,
silent), firing at most one dialog-trigger probe per session while the
decision is pending, and falls back to the previous delivery-probe
heuristics when the helper is unavailable. The card polls the readout
silently in every state, so toggling Allow notifications in System
Settings flips the card within a poll — both directions, verified live.
Test notifications also consult the readout so 'delivered' is no longer
claimed for swallowed notifications.
Packaged builds ship the helper via extraResources and sign it in
afterPack like the computer-use helper; dev copies compile it on demand
(swiftc, non-fatal when missing) with the shared dev bundle id.
Co-authored-by: Orca <help@stably.ai>
* feat: in-app fallback for swallowed notifications + permission card in Settings
- Dispatch now consults the authorization readout before creating a
native notification: when macOS would silently swallow it (denied or
prompt unanswered) it returns reason 'blocked-by-system' instead of
piling invisible notifications into Notification Center. The terminal
notification path surfaces that as a once-per-session in-app toast
with an Open System Settings action. Mobile fan-out is unaffected.
- Settings > Notifications now shows the same live permission card as
onboarding (moved to components/notifications/), polling the readout
so System Settings changes reflect within seconds, and the test
button updates it inline.
- Test sends that are blocked at the OS level now show the
settings-pointing failure toast instead of a generic error.
Co-authored-by: Orca <help@stably.ai>
* fix: hide macOS permission card while Orca notifications are disabled
A green 'Notifications are enabled' card next to a disabled Enable
Notifications toggle read as a contradiction — the card now renders (and
the readout polls) only while Orca's own notifications setting is on.
Also single-flights the authorization helper so simultaneous agent
completions share one readout process.
Co-authored-by: Orca <help@stably.ai>
---------
Co-authored-by: Orca <help@stably.ai>
## Summary
- stop the macOS helper from unlinking caller-supplied socket and token paths
- keep parent-owned token cleanup tied to the helper startup that created it
- fail closed for non-socket socket-path collisions and cover helper cleanup races with regression tests
## Verification
- swift test --package-path native/computer-use-macos
- pnpm exec vitest run --config config/vitest.config.ts src/main/computer/macos-native-provider-client.test.ts
- pnpm run typecheck:node
- pnpm lint
- pnpm run build:computer-macos
- Electron/helper startup smoke validation
* fix(computer-use): stop malformed numeric args from crashing the agent
requiredNumber only checked isFinite, and several request handlers cast the
resulting Double straight to Int/UInt32 (elementIndex, clickCount, pages,
from/toElementIndex, windowId, windowIndex). Int(Double) traps when the value
is outside the integer's representable range, so a single malformed request
such as `{"elementIndex": 1e300}` crashed the entire agent process, killing
all in-flight automation.
Add a bounded conversion helper in the Core library and route every untrusted
Double->integer cast through it. boundedInteger truncates toward zero like
Int(Double) but returns nil (via init(exactly:)) instead of trapping when the
value is non-finite or out of range; the request handlers then surface a clean
invalid_argument error (or resolve to nil for the optional window lookups).
The helper lives in the Core library because the test target cannot import the
executable target where the handlers live. A negative-control run confirms the
out-of-range test traps with the previous Int(Double) cast and passes with the
fix.
* review: close remaining numeric crash paths
- parse stale-element validation indexes through bounded conversion
- reject malformed window selectors instead of dropping the target window
- bound synthetic scroll wheel deltas and cover Int32 conversion
---------
Co-authored-by: Jinjing <6427696+AmethystLiang@users.noreply.github.com>
Neil
OrcaWinandOrca Worker
Jinwoo Hong
Brennan Benson
buf0-bot[bot]andorca-bug-scan-bot