On macOS 26, UNUserNotificationCenter aborts when executables are run
from Contents/Resources because bundleProxyForCurrentProcess returns
nil.
Moving the orca-notification-status helper to Contents/MacOS next to
the main Electron executable ensures proper bundle resolution and
avoids immediate crashes.
* auth v1
* fable review
* lint
* Account menu with org membership management
Default UX is a compact account menu (sign in, organization selection, sign
out) that renders only when cloud auth is configured; adds an organization
members dialog (invite, role, remove) gated on server-side role checks. The
multi-profile switcher UI is preserved behind ORCA_MULTI_PROFILE_UI=1.
Co-authored-by: Orca <help@stably.ai>
* Gate the optional account sign-in UI to dev builds
The account switcher stays hidden in packaged builds while the feature is
in progress. Dev builds still show it when the client env vars are set, and
a dev-only Settings > Dev Tools > Orca Cloud section mirrors it.
Co-authored-by: Orca <help@stably.ai>
---------
Co-authored-by: Orca <help@stably.ai>
Two onboarding-screen bugs:
- The final "notifications" step blocked click-off/Escape dismissal, unlike
every other step. Remove the notifications-only guard so the skip
confirmation opens on all steps; the footer "Skip to project setup" stays
hidden there since the primary button already hands off to Add Project.
- A skipped "integrations" step (GitHub CLI already installed) still rendered
as a dead, disabled stepper dot the user skipped past on Continue. The
stepper now drops all skipped steps (integrations + Windows terminal)
entirely instead of showing an unreachable dot.
Allowing dismissal on the last step let a click-off race the "Add your first
project" completion handoff (both call closeWith) and double-write onboarding
state / double-fire telemetry. Make closeWith idempotent with a first-wins
latch. Also map the displayed step index through resolveStepIndex so a
momentarily-skipped resume step can't flash "1 of N".
Verified: onboarding unit tests, full onboarding e2e spec (rewritten
notifications test locks in the new dismiss behavior), typecheck, lint, and
live Electron.
The warm-reattach mouse-leak e2e (added in #7893) awaited the readiness
marker unconditionally, so it hard-failed on sandboxed runners whose PTY
echoes input but never execs a shell. Wrap the gate in a skip guard that
mirrors the existing pane-manager guard: run the assertions where the
shell executes, and skip gracefully otherwise instead of a false red.
Co-authored-by: Orca <help@stably.ai>
* 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>
* fix(terminal): clear leaked mouse-reporting modes on pane reattach
A TUI that enables mouse tracking (?1000/1002/1003 + SGR 1006/1016) and
dies uncleanly never emits the disable sequence, so the daemon's snapshot
records the mode and buildRehydrateSequences re-arms it on every reattach.
POST_REPLAY_REATTACH_RESET cleared cursor/focus/kitty state but not mouse
modes, so a plain shell in the reattached pane echoed every pointer-motion
report (`<35;col;rowM`) as literal text.
Add RESET_MOUSE_REPORTING (?9l ?1000l ?1002l ?1003l ?1006l ?1016l) to both
POST_REPLAY_REATTACH_RESET and POST_REPLAY_MODE_RESET. Live agent panes keep
mouse modes via POST_REPLAY_LIVE_AGENT_REATTACH_RESET, so agent scroll is
unaffected.
Verified against the real daemon serializer + real xterm: the old reset
leaves mouseTrackingMode armed, the new reset returns it to 'none'.
* test: add regression tests for terminal mouse mode leak on reattach
- Add an E2E test to verify that a warm reattach disarms mouse modes
left armed by an uncleanly exited TUI.
- Add a test fixture that writes the mouse tracking enable sequence
without a matching disable sequence.
- Ensure the reattached pane disarms mouse tracking and that actual
mouse movement produces no reports.
* Add types to isMouseReport in terminal reattach leak test
Explicitly annotate parameter and return types for the isMouseReport
helper function inside the page.evaluate block.
* Refactor mouse-mode leak E2E test to use shell printf and live pane
Eliminate the external Node.js fixture file and streamline the E2E test
by using a POSIX printf shell builtin to arm the mouse tracking modes.
Additionally, verify the leak precondition by inspecting the active pane's
live terminal state (mouseTrackingMode) rather than querying internal
daemon buffer snapshots via window.api.pty.getMainBufferSnapshot.
Replace the sidebar layout in the GitHub issue details view with a
responsive grid of top columns placed above the description body. This
ensures the description content is not squeezed by a right rail and fits
the header's full content width.
Identify Claude sessions with no saved conversation turns but possessing
recoverable signals such as queued prompts or subagent transcripts.
This displays them in the sidebar with a "Not saved" badge instead of
filtering them out as empty, and provides detailed notices to recover
them via logs while disabling standard resume actions.
Additionally, extract Gemini session parsing logic into a separate
module and support counting sibling subagent transcripts across local
and remote SSH session scans.
Show a small amber dot on the floating-workspace launcher (both the
floating-button and status-bar triggers) whenever any floating-workspace tab
still has an unacknowledged terminal bell or agent completion, and a
composited amber dot on the Windows tray icon when the window is
minimized/hidden. Both clear through the existing show-until-interact paths —
engaging with or closing the offending tab drops the dot with no stale unread
state left behind.
The launcher dot derives from the existing per-tab/per-pane unread maps via a
new selectFloatingWorkspaceHasUnread selector (primitive boolean, empty-
workspace early return, no bespoke state). The tray dot rides the notification
dispatch and clears on window show/restore.
* docs: design fix for sticky OPEN PR after merge
Capture root cause and primary fix for Checks panel preserving open/draft
PR cache on authoritative no-pr after merge + HEAD diverge.
* Clear open and draft PR caches on fallback refresh misses
Avoids preserving non-terminal ("open" or "draft") PR states in both
the PR and hosted-review caches when a fallback refresh returns an
authoritative "no-pr" result. This resolves a sticky "OPEN" UI bug
where a merged PR continued to show as open.
- Removes fallback PR preservation from the main PR cache check.
- Gates the hosted-review cache fallback preservation to only accept
terminal states ("closed" or "merged").
- Updates unit tests to assert cache clearing for open/draft states.
* Remove stale open PR refresh design document
Delete the design and diagnosis document for the stale open PR checks
panel refresh issue now that the investigation and planning phase is
complete.
- Keep the explicit "Push" and "Force Push" dropdown options enabled when
there is no upstream or the upstream status is loading, letting Git
resolve the target at execution time.
- Stop auto-upgrading regular push operations to force push in the background,
ensuring normal push actions stay non-force.
- Explicitly route the primary action and commit-and-push flows to the
"force_push" action when a lease force push is required.
* fix(mobile): normalize Grok rotating OSC titles at the main observation boundary
Desktop already collapses Grok Build's rotating working frames via the
renderer's normalizeTerminalTitle, but the main process stored raw OSC
titles, so mobile session tabs (fed from pty.lastOscTitle) still saw a
distinct title every spinner frame and re-touched snapshots each time.
Apply normalizeTerminalTitle once where main records an observed OSC
title, before the prevTitle comparison that gates session-tab and
mobile-snapshot touches, and normalize hydration-seeded titles the same
way so the first live frame after a seed compares equal. Agent status
stays detected from the raw title, mirroring the renderer tracker.
Covers remoted/SSH PTYs too since they surface through the same main
observation path.
Co-authored-by: Orca <help@stably.ai>
* Fix Grok and Pi terminal title normalization and status detection
- Require a strict "spinner - phrase - grok" pattern for Grok working
titles to prevent false positives on other agent tasks ending in
"- grok" (such as Claude or Codex).
- Treat collapsed "Pi" and "OMP" synthetic titles as idle by default
to prevent them from reverting to neutral status after normalization.
- Prevent duplicate mobile session tab updates during Grok status
frame rotations by ensuring they normalize to a stable title.
---------
Co-authored-by: Orca <help@stably.ai>
Introduce an external link action in the header of the Mobile PR View
Panel. This provides a persistent and easily accessible shortcut to open the
current pull request's canonical URL in the system browser.
Cursor Agent parks the real cursor on a blank row and hides it (?25l),
drawing its own caret. Orca's post-replay resets forced ?25h, painting a
stray cursor block below the agent's prompt after reattach and after
hidden-to-visible snapshot restores.
Root cause fix: stop re-showing a cursor the replayed bytes intentionally
hid, instead of compensating afterwards.
- Live-agent reattach reset now preserves the payload's final DECTCEM
state (last ?25l/?25h wins); the post-parse viewport veto re-shows the
cursor when the screen disproves a live parked agent, so a shell can
never inherit a permanently hidden cursor from a dead TUI's leftovers.
- Hidden-output restore now uses a live-agent variant of the snapshot
reset when status/title corroborates a live agent: forcing ?25h
re-showed the parked cursor, and ?1004l permanently silenced the
focus-in the agent needs to unpark (agents only enable focus reporting
at startup, so nothing ever re-armed it).
Co-authored-by: Orca <help@stably.ai>
* feat(source-control): add Fix push failure with AI for pre-push hooks
Detect pre-push hook failures separately from auth/transport errors so
push toasts and inline messages no longer suggest checking repo access.
Mirror the commit-failure recovery flow with a fixPushFailure action,
summary panel, details dialog, and agent launch recipe in Settings.
Fixes#6497
* feat(source-control): add push failure AI recovery
Co-authored-by: dalveytech-vincent <vincent@dalveytech.com>
---------
Co-authored-by: dalveytech-vincent <vincent@dalveytech.com>
Grok Build's working OSC titles interpolate a rotating status or tool
phrase between the spinner and its name, causing tab and sidebar
titles to fluctuate rapidly. Collapse these working frames to a stable
"⠋ Grok" label, while leaving idle and session titles untouched.
* Show empty sessions by default in AI Vault Panel
Change the default state of `hideEmptySessions` from true to false, and update the active filter count calculation to treat false as the new default.
* Centralize AI Vault view defaults and fix reset state alignment
Consolidate the default values for sorting, grouping, and hiding empty
sessions into a single place. This prevents inconsistencies between the
initial state, the badge count adjustment checks, and the "Reset" view
action, where the reset previously toggled hideEmptySessions to true
contrary to the initial false value.
* Count agent adjustments by membership instead of list length
Ensure that swapping one agent for another (which keeps the total
array length the same) is correctly identified as a deviation
from the default of having all agents enabled.
Claude's AskUserQuestion tool is auto-allowed, so it emits a PreToolUse (not
PermissionRequest) while blocked on a human answer. normalizeClaudeEvent mapped
that to `working`, so the sidebar showed a spinner that decayed to grey while
the question sat. Map the auto-allowed AskUserQuestion PreToolUse to `waiting`
instead, mirroring the Kimi/OpenCode handling, so the row and worktree dots read
amber "Waiting for input".
Separately, a frozen braille-spinner title left by an exited agent (e.g.
"⠐ Review branch for regressions" over a shell prompt) kept classifyTitleActivity
returning `working`, spinning the worktree dot forever with 0 agents — the row
builder rejects the same unattributable title, so no agent row is shown. Gate
tabHasStatus's title-derived working/permission on the same agent attribution the
row builder uses (resolveAgentTypeFromTerminalTitle), so a title only spins the
dot when it would also show a row. Hook-driven status (hasLiveWorking/
hasPermission) is untouched.
* fix(terminal): shape Arabic/RTL runs on the cell grid via xterm character joiner (#5262)
Register a character joiner that groups contiguous RTL runs (tunneling
through neutral spaces/digits/punctuation between RTL words) so both the
WebGL and DOM renderers draw each run as one string, applying native
contextual shaping and BiDi ordering inside the run's grid-aligned cell
box. Buffer and PTY stream are untouched; xterm un-joins ranges holding
the cursor or a partial selection, keeping cursor visibility and
selection cell-accurate. Non-RTL lines early-return via a single
charCodeAt sweep.
Co-authored-by: Orca <help@stably.ai>
* Deregister Arabic shaping joiner when disposing pane
Since xterm.js does not automatically remove registered character joiners
upon terminal disposal, they must be explicitly deregistered to avoid
memory and registration leaks.
- Update `registerArabicShapingJoiner` to return a cleanup function.
- Store and call the cleanup handler when disposing a managed pane.
- Update tests to assert proper registration and deregistration.
* fix(terminal): gate RTL joining on live WebGL, tunnel ZWNJ/ZWJ/RLM, lock joiner lifecycle in tests
Review findings on the Arabic shaping branch:
- The DOM renderer sizes a joined span with one letter-spacing value applied
per character, so joined RTL runs blew out row grid alignment whenever the
WebGL renderer was unavailable (GPU off, auto policy, context loss). The
joiner now only returns ranges while the pane's WebGL addon is live.
- ZWNJ/ZWJ/RLM (mandatory in Persian orthography) hard-broke RTL runs,
rendering word halves in swapped visual order. They are now transparent
inside a run without extending or counting toward it.
- openTerminal's existing unicode-ordering test failed at HEAD because the
fake terminal lacked registerCharacterJoiner; the harness now implements
the joiner API and a new test locks register-on-open/deregister-on-dispose.
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): keep zero-width RTL marks from opening joined runs, treat ALM as transparent
Round-2 review findings:
- A combining mark (tashkeel/niqqud) orphaned after an LTR base could open
an RTL run mid-cell; xterm rounds that string range to an empty joined
cell range and the WebGL renderer draws an empty glyph over the next
character. Marks now extend and count only in runs opened by a spacing
RTL letter.
- ALM (U+061C) sat inside the strong-RTL block scan, so it opened and
counted toward runs unlike RLM. It now joins the transparent set.
- New test locks openTerminal wiring the joiner predicate to
pane.webglAddon (previously () => true would pass the suite).
Co-authored-by: Orca <help@stably.ai>
* fix(terminal): treat zero-width Cf controls in RTL blocks as run-transparent
Round-3 review finding: zero-width format controls that live inside the
strong-RTL scan ranges (Arabic number signs U+0600-0605, end of ayah U+06DD,
Syriac abbreviation mark U+070F, disputed end of ayah U+08E2) and the BOM/
ZWNBSP U+FEFF are width-0 in xterm. Like combining marks they could open a
run mid-cell, producing an empty joined cell range that blanks the following
glyph in the WebGL renderer. They are Cf not Mn, so canOpenRtlRun could not
catch them. Fold them into the run-transparent set (never open, count, or
break a run) and document the upstream standalone-width-0-cell join skew.
Co-authored-by: Orca <help@stably.ai>
* Document WebGL decoration limitation for Arabic shaping joiner
Add comments explaining a known upstream limitation in xterm.js WebGL
renderer where search-match highlights inside a joined run render
all-or-nothing, similar to ligatures.
---------
Co-authored-by: Orca <help@stably.ai>
* feat(mobile): add explicit keyboard dismiss control to terminal command dock
Add a fixed Hide control at the left of the terminal command dock accessory
bar whenever the software keyboard is open (keyboardHeight > 0). Tapping it
clears any pending live-input focus timer, blurs the live and buffered command
inputs, and dismisses the keyboard without sending bytes, switching input mode,
or clearing typed text.
The dismiss behavior lives in a dedicated, unit-tested terminal-keyboard-dismiss
module rather than the customizable accessory-key path, so the escape hatch
cannot be hidden by user shortcut customization. Available on every platform
where the IME covers the app (iOS and Android).
* review: harden keyboard dismiss control per adversarial review
- document the load-bearing clear-before-blur order in dismissTerminalKeyboard
- cover the both-handles-missing case in unit tests (5/5)
- move the #5106 first-tap comment onto the accessory ScrollView and add a
why-comment for the fixed Hide control
- add accessibilityRole=button and hitSlop to the Hide control for a larger,
semantically-correct touch target
* fix(mobile): harden hide button visibility and scroll layout
* refactor(mobile): use stacked keyboard+chevron glyph for dismiss control
Replace the icon+'Hide' text with the iOS-native dismiss glyph (keyboard
with a chevron-down beneath it). Narrower in the accessory row, removes the
icon/word redundancy, and reads as distinct from the >> input-mode toggle.
Accessibility label/hint/role unchanged.
* fix(mobile): align keyboard dismiss accessory height
* test(mobile): align vitest transform with Vite 8
---------
Co-authored-by: Wolfgang Schoenberger <221313372+wolfiesch@users.noreply.github.com>
Co-authored-by: Jinwoo-H <jinwoo0825@gmail.com>
After #7750 removed the containment guard, the residual #7239 failure mode
is a persisted/inherited startupCwd whose directory no longer exists: every
spawn dies with the provider's missing-directory error. Fresh local renderer
spawns now opt in (cwdFallback: 'worktree') to recover at the workspace root
with a generic in-terminal notice; reattach, SSH, remote-runtime, runtime/API
and mobile callers keep exact cwd semantics, and existing directories —
including outside the worktree (#7685) — spawn as requested.
Co-authored-by: Orca <help@stably.ai>
Brennan Benson
OrcaWin
github-actions[bot]
Jinjing
Neil
Jinwoo Hong
Wolfie
Rod Boev