Commit Graph
20927 Commits
Author SHA1 Message Date
Jinjing 5f3b2fa4bf feat(artifacts): allow artifact publishing by default for new profiles (#40948)
New profiles default artifactSharingEnabled to true. Existing users are
unchanged: stored settings win over defaults on load, and a stored profile
saved before the key existed loads as false instead of inheriting the new
default. The gate still requires an explicit true, sign-in is still
required, and web clients still mirror the host (their local fallback no
longer inherits the desktop default). showArtifactsButton is unchanged.
2026-11-10 17:28:29 -08:00
c603a3617f fix(github): show Actions jobs as checks when a token cannot read check runs (#39843)
* fix(github): show Actions jobs as checks when a token cannot read check runs

A fine-grained personal access token cannot be granted the Checks
permission, so GitHub refuses every check-run read and the checks panel
failed with 'Resource not accessible by personal access token'. When the
REST check-run read is refused that way, the fallback now lists the head
commit's GitHub Actions jobs (Actions: read) in their place; a job's id is
its check run's id, so job logs keep working.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(github): read Actions jobs in the shared REST check pages

Moves the fine-grained token fallback into readRestPRCheckRuns and
readRestPRCheckSuites so it reuses the existing pagination and also
covers the CI summary and PR watch reads. Check details skip the
refused check-run read and still load the workflow jobs and logs.

* fix(github): read refused check details from the Actions job

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Kelvin Amoaba <97001695+AmoabaKelvin@users.noreply.github.com>
2026-11-11 01:25:56 +00:00
OrcaWinandOrcaWin dcc21242d9 test: remove 93 low-value cases and eleven suites (#40957)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-10 17:23:31 -08:00
6575920247 fix(settings): list the host's models in agent settings (#38764)
* fix(settings): list the host's models in agent settings

Settings read the built-in model list, so new Codex models never appeared.
It now reads agentSession.modelCatalog when an agent's row is open.

Fixes #38643

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(settings): mark the built-in model list and keep new-chat defaults out

Say when Settings shows the built-in list because the host gave none.
Read the host's answer through the snapshot, which drops a listed default
that a workspace's own config could replace.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(settings): show Codex's account models and don't trust an old list

A Codex account answer is scoped by workspace, so reading it through the
new-chat snapshot brought back the built-in list. Use the answer directly.
A failed read now shows the marked built-in list, not an older answer.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(settings): build the older model list without a type assertion

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(settings): read the model list again when the agent command or account changes

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(settings): build the changed settings without a type assertion

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(settings): assert the shown models instead of the read count

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* test(settings): restore the store settings after each model list test

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(settings): refresh the model list when a host drops its lists

A paired host can change account on reconnect with no setting changing here,
so an open row kept the old list until it was reopened.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(settings): follow the moved native-chat runtime modules

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(settings): show the built-in list note only when there is a built-in list

* fix(settings): keep the Settings model list current while a host is still listing

Wait again when a host listing outlasts one wait, re-read when a paired host's launch
settings change, and stop writing the account answer into the new-chat snapshot, where
it dropped newer workspace answers.

* fix(settings): ask paired hosts to wait only after they report a running listing

Older hosts refuse waitForListing, so the first read is plain, the same order the new-chat
picker uses. A paired host's settings change re-reads without hiding the list shown.

* refactor(settings): read the host catalog without a hooks-rule suppression

---------

Co-authored-by: temp_final_v5 <>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Kelvin Amoaba <97001695+AmoabaKelvin@users.noreply.github.com>
2026-11-11 01:20:37 +00:00
Kelvin Amoabaandiskic-marko 6009e6423e fix(droid): keep the mission's session as the tab's resume target when sub-droids report (#40947)
* fix(droid): keep sub-droid sessions from replacing the pane resume target

Mission workers inherit the orchestrator's pane env, so their hooks
reported their own session ids on the orchestrator's pane and the last
worker became the tab's resume target after a restart.

Fixes #39278

* fix(droid): keep the orchestrator's session after the pane row is retained

Count a retained Droid row as the pane's owner, so worker hooks that arrive after the row is dismissed or reseeded still keep the orchestrator's session, and never hand a worker's own id to a pane with no Droid owner. Moves the test next to the other provider server tests.

* fix(droid): keep the pane's Droid session when a relay forwards a worker hook without one

---------

Co-authored-by: iskic-marko <142354657+iskic-marko@users.noreply.github.com>
2026-11-11 01:18:03 +00:00
Neil 8aa3f9c36e refactor(renderer): group browser owners and typing diagnostics (#40936) 2026-11-10 17:17:59 -08:00
Brennan Benson 985176b11b fix(remote-runtime): reopen client-event streams after connection replacement (#40881)
* fix(remote-runtime): end remote streams when the host connection is replaced

A remote stream remembered its host connection's transport generation and,
once the generation moved on, main silently dropped every frame without
closing the stream. The chat reader still believed it was connected and never
reopened, so messages stayed "Sending…" until the tab was reopened.

Advancing the generation (including eviction) now ends every pending and
retained subscription fenced on the old one; readers resume from their applied
cursor on the new connection.

Tracked in STA-14427.

* fix(remote-runtime): resubscribe client events when their stream ends

Main now ends streams whose host connection was replaced, but the renderer's
client-event owner ignored the close and kept a dead subscription while its key
stayed the same, so repository, worktree and SSH updates stopped. It now drops
that subscription on close (also when the close lands mid-setup) and
resubscribes through its existing retry.

Tracked in STA-14427.

* fix(remote-runtime): back off client-event resubscribes until a stream stays up

A host that accepted the client-event subscription and then dropped it made
the owner resubscribe about once a second forever, because each install reset
the failure count. Every close now counts toward the existing capped backoff,
and the count resets only once a stream has stayed up for 30 seconds.

Tracked in STA-14427.

* fix(remote-runtime): retain the landed transport retirement implementation

* fix(remote-runtime): isolate client-event retries by host
2026-11-10 17:08:11 -08:00
Kelvin Amoaba 2259dbe163 fix(status-bar): import window visibility interval from its new path (#40951) 2026-11-11 01:07:50 +00:00
Brennan Benson bc35323f17 Keep Chat UI model selections from snapping back to older replies (#40906)
* fix(native-chat): keep a model pick shown instead of an older answer's model

A saved model pick on a live structured Claude chat snapped back: the host's
options answer kept returning the model that answered before the pick, and the
picker shows a reported model over the selection (for provider fallback).

A pick now retires the child's earlier model reports until a turn carries that
pick, as the common pattern does with a fallback report once the user picks
again. The Claude adapter ignores old-execution frames until a turn applies the
picked value; the host drops a report captured by an options read when a pick or
a child change lands before the read answers. Picks still reach Claude at the
next turn.

* test(native-chat): cover the confirmed-init report under read freshness

* test(native-chat): cover picks during recovery and legacy no-op mutations
2026-11-10 17:06:19 -08:00
Neil 3d9be7c269 relay: an edge fence refused as a same-epoch conflict goes once more one epoch above (#40946)
At 1% the edge answered fences 'conflict': it holds the fence's own epoch for another placement
(a move onto the edge whose push landed though the director counted it failed, then a placement
elsewhere minted the same epoch). Retries go to the host's current row at that same epoch, so they
kept failing until the host moved again; meanwhile the edge object claims the host and its phones
get HOST_OFFLINE instead of re-resolving. No later move can produce a same-epoch conflict (it would
answer stale), so the fence now goes once more one epoch above, which lands; a later move onto the
edge re-mints above it. Stale answers are never bumped.
2026-11-10 16:58:38 -08:00
OrcaWinandOrcaWin 049c0a7421 test: remove eight low-value suites and routine renderer cases (#40941)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-10 16:52:02 -08:00
Brennan BensonandClaude Opus 5.5 8ac9fb6eee Stop re-reading every linked GitHub issue when the window returns (#40534)
* fix(github): stop re-reading every linked issue when the window returns

Each window reveal re-read the linked GitHub issue of every workspace, shown or not,
through the shared 4-slot gh queue, so PR, Checks and Merge waited behind hundreds of
issue reads, and above 500 linked issues the cache thrashed on every reveal.

Surfaces that show an issue now own its freshness: the reveal handler no longer sweeps
the catalog; workspace cards keep their own visible-only issue poll, now also in the
paired web client and for folder workspaces; the space manager's decision hover polls
its one issue while open.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* fix(space): the open space-manager hover asks the main window to refresh its issue

The desktop Space window shows the main window's published issue cache, so a read done
in the Space window was overwritten by the next snapshot. The decision hover now owns
its issue's freshness only while open, follows the live link, and asks the main window
(the cache owner) to re-read it; the web client reads in place.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* fix(space): route the hover's issue refresh by window and follow owner changes

Choose the main-window relay only in the separate Space window (the paired web client's
fallback API made the relay look available), and refresh when the worktree's owning host
changes, not only its link.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* perf(space): key the open hover's issue refresh off the row projection

The hover re-resolved its worktree owner from the store on every Space-window snapshot,
rebuilding the catalog lookup index each time. Use the cache key the row projection
already computes for the issue it shows.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* test(space): use typed fixtures in the hover freshness test

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-11-10 16:46:16 -08:00
4a31139c50 fix(pty): scope Linux terminal foreground rows to the anchor session (#39557)
* fix(pty): scope terminal foreground rows to the anchor session on Linux

A docker run -t container holds its own private pts/N, and Linux reports
terminal membership only by number, so ps -t pts/N listed the container's
processes beside the agent's. The foreground coherence check then saw a
tpgid that never matched and refused guarded writes with
agent_not_in_foreground (#39488).

Prefix the Linux terminal ps with a sid column (macOS ps has no sid
keyword) and keep only rows in the anchor process's session; a missing
anchor row keeps every row as before.

Co-Authored-By: Claude Code <noreply@anthropic.com>

* fix(pty): require the anchor PID for terminal rows and test the real ps shape

The session filter is the fix, so every caller now passes the anchor. The
regression test answers in whatever columns ps was asked for, so reverting the
fix fails on the foreground coherence check rather than on a parse quirk.

---------

Co-authored-by: Claude Code <noreply@anthropic.com>
Co-authored-by: Kelvin Amoaba <97001695+AmoabaKelvin@users.noreply.github.com>
2026-11-11 00:46:03 +00:00
Kelvin Amoaba c58ffc0042 revert(mobile): restore the scrolling session tab strip (#40938)
Reverts #40592 (12e809c755). The session bar goes back to the scrolling strip of tab cells, the + menu goes back to the action sheet, and the header subtitle shows the tab count again.
2026-11-11 00:41:08 +00:00
Jinjing a7a21c8658 Source Control menu: one line per row with a short status (#40919)
* Show Source Control menu rows on one line with a short status

* Keep Push and Force Push clickable; force push shows remote-only commits it replaces

* Match Force Push description to remote-only commits; list review blockers exhaustively

* Show short statuses for every blocked review action
2026-11-10 16:40:17 -08:00
buf0-bot[bot] 06f97dee35 Update README downloads badge (#40940)
Co-authored-by: buf0-bot[bot] <252831055+buf0-bot[bot]@users.noreply.github.com>
2026-11-11 00:39:37 +00:00
Brennan Benson fd20d34362 Save the issue title on GitHub issue links made without one (#40853)
* fix(github): title GitHub issue links that were saved without one

A workspace linked to a GitHub issue from the CLI (`--issue N`), a phone, the paired web
client or a typed number kept only the number, so search could match its issue title only
while a client happened to hold that issue. The host that owns the workspace now reads each
newly linked untitled issue once, after the write returns, the way its card would, and saves
the title on the link. It keeps the current review selection and never resurrects a removed link.

* fix(github): bound and publish issue-title fills like other metadata writes

Skip writes that leave links untouched before normalizing, check a queued link is still wanted
before reading it, publish filled titles once per repo per 2 s window, refresh the resolved rows
remote clients read, never wake a stopped WSL distro for a background read, keep a selected
review's PR suppression, and keep a failing write observer from failing a saved write.

* fix(github): keep link-title fills off unrelated writes and out of per-title refreshes

Ignore linkedWorkItem in the unchanged-links check (it is rebuilt on every write and always
titled), publish filled titles at once on a quiet host and at most once per repo per 30 s
during a burst on a timer, and only lift a PR suppression when the selected PR changes, so
restating the current PR (as a title fill does) keeps a hidden branch PR hidden.

* perf(github): publish filled link titles as addressed workspace metadata

Instead of a whole-repo worktreesChanged refresh per burst window, the host publishes one
addressed workspaceMetadata set record {hostId, epoch, seq, id, row} for the workspace whose
link was titled: to its desktop window over IPC and to paired clients as a runtime client event.
Clients patch that one row (Cmd-J reads it) and drop replayed records; older clients ignore the
new event and see the title on their next listing.

* fix(github): agreed workspaceMetadata record shape, per-id seqs and late-listing refill

Records name their kind and domain, clients keep the last applied seq per workspace for each
producer epoch (dropped when the epoch changes), and a listing that read metadata before the
title write cannot drop the title: recorded titles refill rows that still hold the same untitled
link for 60 s, never restoring a removed link. The IPC channel is marked as interim transport.

* fix(github): title-only, host-addressed application of workspaceMetadata records

Clients fill the record's issue titles into the named row's still-untitled links and never take
its links over the row's current ones, so a record that arrives after an unlink cannot restore
the issue. A paired host's records reach the row of their own execution host (an SSH twin is a
different row), and go through the bridge's connection-generation check like other host events.

* test(github): type the metadata-record test helper as AppState

* fix(github): keep recorded titles until expiry, refill only changed repos, retire with the pairing

A record that beats the listing adding its link keeps its titles for the full window; the refill
runs only for repos whose listing changed, indexes that repo's rows once and skips rows already
checked; per-workspace seqs live in the expiring entries; a re-pair or reconnect of the producing
environment retires its titles.

* fix(github): read the listing through the workspace record selector

* fix(github): keep link titles until their producer changes; fall back to a listing on eviction

Recorded issue titles no longer expire on a timer (a slow listing could outlast it): they live until
the producer's epoch, pairing or connection changes, refilling only present untitled links. When
more than the cap are in flight the oldest is evicted and its repo is listed afresh, which carries
the saved title. The refill runs in updater form so it never merges a stale store snapshot.

* fix(github): sequenced listings replace the retained-title refill

Per the re-plan ruling: the store keeps one workspaceMetadata cursor per process; a title record takes
the next seq in the same turn as its write, and each listing records the cursor when it reads metadata.
Clients note each record's seq per producer and repo; a listing at or past it ends the obligation, an
older one keeps titles the rows already show and lists the repo afresh (a new invocation, never joining
the stale one), a record whose row or link isn't listed yet asks for a fresh listing, and a seq gap
lists every repo of that producer. The retained-title table, its watcher, TTL and cap are gone.

* test(github): stamp capture, gap, restart and recovery coverage for workspace metadata listings

Also retire an environment's earlier coverage bookkeeping when it is re-paired or reconnects.

* fix(space): keep issue-title coverage per repo and host until a newer listing lands

Coverage now keeps each projection's newest record seq after its covering listing settles, so an older listing that finishes later still keeps shown titles. Projections are keyed by repo and host, paired producers keep their stream cursor, retired pairings and unlisted projections are dropped, every listing served by this host carries its metadata stamp, and a record with any unlisted titled link asks for a listing.


* fix(space): re-list after any stale listing and drop retired metadata producers

A stale listing now always asks for one fresh listing, so a link it dropped comes back and a failed scan can no longer end the wait. Removed environments drop their producer state, the producer's repo list is built only on a gap or a first-seen projection, and recovery listings skip the lineage refresh.


* perf(space): filter current rows only for listings that keep titles


* fix(space): forget removed repos on a metadata gap; keep titles for another producer run's listing


* fix(palette): search a linked issue by its saved title before its card renders

The issue's search entry took its title only from the issue cache, which fills when the card renders, and it shadowed the saved link's titled entry under the same id. It now falls back to the link's saved title.


* perf(palette): read a row's links once per palette document


* fix(palette): give each saved link its own search unit

Saved-link evidence was meant to carry an `attachment:` id but the override set an unused field, so a link shared the `issue:N` unit id with the legacy issue field and with same-number links from other repos, and only the first (often untitled) survived. The id now applies to the unit and its fields, so every saved link's title is searchable. Replaces the cache-title fallback.


* fix(space): keep the title queue draining past a failed link; encode environment ids in coverage keys


---------
2026-11-10 16:38:44 -08:00
Neil 08bb458da4 Remove the per-repo agent folder trust prompt (#40934)
The first-launch 'Trust <repo> for agents?' modal blocked agent launches
to ask a question that offered no real protection: undecided repos, and
every unattended launch, already pre-trusted. Drop the dialog, the launch
hold and the skipAgentWorkspaceTrust plumbing. Repos default to
pre-trust; the per-repo setting remains as an opt-out.
2026-11-11 00:34:24 +00:00
Kelvin Amoaba c20f528d3e fix(omp): show the picked model's own thinking levels before the first chat message (#40933)
OMP's thinking menu belongs to the running model, but Orca copied it onto
every other OMP model in a new chat. Treat OMP like OpenCode: leave
non-running models' menus unknown, drop such copies saved by older
builds, and read a picked model's menu from a throwaway listing session
started on that model. Starting on it, rather than switching to it,
keeps OMP from recording the model as used in the user's account; if
OMP cannot start on it, the listing runs again without it.

A read that asks for a picked model's menu now reports the listing in
progress, so an open picker updates when the menu lands instead of on
the next open.

Fixes #39015
2026-11-11 00:27:08 +00:00
OrcaWinandOrcaWin 48d5bfd76c test: remove four low-value renderer suites and redundant cases (#40928)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-10 16:26:37 -08:00
Neil 7fa124de5e relay edge: silent host at authorization expiry closes 4408; dead-control diagnostics (#40929)
* relay edge: a keepalive host already silent when its authorization expires closes 4408

At token exp + 60 s, a keepalive control that went silent (asleep, offline) had its silence probe
pending, so the expiry check closed it 4401. Production showed these as 4401 'never refreshed'
closes with the host silent for ~5-6 minutes, inflating the auth-failure count and hiding drops
from the 4408 comparison with the cells. Such a host now closes 4408 (orphan time back-dated as
for any silence close). Local e2e K4 (fails without the change: 4401). 48/48.

* relay edge: handshake time on control_active, control age and silence on dead-control closes

At 1%, many new controls went silent at or right after activation and were closed minutes later,
with no rebound from a retry. To tell a desktop that gave up the connect (15 s deadline, e.g. a
slow first hello on a new object) from one that died later: control_active double2 = ms from the
upgrade to activation; new silence_close (seconds silent, where found, keepalive/legacy, age);
auth_expired_close double3 = control age.
2026-11-10 16:19:45 -08:00
Neil 1af64194a4 refactor: group renderer agent sessions and runtime modules by owner (#40911)
* refactor: group renderer agent sessions and runtime modules by owner

* fix: carry moved owner paths into refactor ratchets

* test: remove retired empty sleep guard registration
2026-11-10 16:04:03 -08:00
OrcaWinandOrcaWin 6192c9b8d5 test: remove six low-value suites and narrow redundant cases (#40922)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-10 15:55:48 -08:00
c3662026f5 Fix workspace delete shortcut targeting (#40913)
* Fix workspace delete shortcut targeting

Resolve the keyboard action from the current workspace and execution host,
keep clicked menu actions bound to their rows, and advertise the shortcut
only for the committed workspace copy. Preserve confirmation preferences.

Co-authored-by: Luca Critelli <lucacri@gmail.com>
Co-authored-by: Anton Kuligin <anton.kuligin@notraffic.tech>

* Document isolated deletion test fixture boundaries

* Type the shortcut command test input

---------

Co-authored-by: Luca Critelli <lucacri@gmail.com>
Co-authored-by: Anton Kuligin <anton.kuligin@notraffic.tech>
2026-11-10 15:51:05 -08:00
Neil 8b02c184bb fix(claude): only offer theme sync when Claude Code CLI runs in a terminal (#40916)
Native chat (agent-session) tabs project Claude status entries into
agentStatusByPaneKey, which made the terminal-agent presence check fire
the Claude Code theme-sync offer without the CLI ever being launched.
Status entries now count only when their pane belongs to a terminal tab.
2026-11-10 23:39:39 +00:00
OrcaWinandOrcaWin 0db5747634 test: retire low-value renderer projections and unused HTTP fixture (#40915)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-10 15:36:37 -08:00
Kelvin Amoaba 9cedbae881 fix(native-chat): browser run's Show button opens its own chat's page, not another agent's (#40910)
* fix(native-chat): show the page a browser run drove, not another chat's

* fix(browser): run each queued agent command in its own RPC call

* fix(native-chat): once a chat claims its browser activity, ignore unclaimed activity
2026-11-10 23:32:57 +00:00
Kelvin Amoaba c4e4428af7 fix(mobile): say which WebView to install when Android's is too old to open a workspace (#40907)
* fix(mobile): tell Android users which WebView to update when it is too old

The page shell needs androidx's DOCUMENT_START_SCRIPT:1, which stable
WebView only reports from Chromium 101; the comments claimed 83/88.
Name the required WebView and the one installed, and how to switch it.

* fix(mobile): read the WebView provider through the shell module index

Tests stub the module index; a direct file import loaded expo-modules-core unstubbed.
2026-11-10 23:32:37 +00:00
Brennan Benson bc1427f6ce Keep repositories on their known public forge (#40542)
* fix(git): keep public forge ownership ahead of auth evidence

* test(git): exercise forge admission with complete auth fixtures
2026-11-10 15:26:25 -08:00
Brennan Benson 8d45e46743 fix(claude): preserve each account's runtime state during refresh (#40558)
* fix(claude): preserve profile-owned state during configuration refresh

* test(claude): bound profile state publications at workspace scale

* fix(claude): honor profile API approval before shared rejection
2026-11-10 15:26:00 -08:00
Brennan Benson a775acc780 Keep loading Git Graph commits until a tall pane can scroll (#40884)
* fix(git-graph): keep loading commits while the end of history is in view

Git Graph only asked for the next page from a scroll event, so a pane
taller than the first 50 commits (or one enlarged later) had no scrollbar
and never loaded older history. The next page is now requested whenever
the end of the loaded rows is within the visible range, once per loaded
result; a failed page waits for Load more or Refresh, and a page that adds
no commits keeps the previous result so it cannot repeat.

Tracked in STA-14490.

* fix(git-graph): keep viewport paging's once-per-result latch across remounts without retaining old history

* fix(git-graph): page only the result whose rows were measured in view

* test(git-graph): replaced history is released while the pane stays mounted
2026-11-10 15:25:16 -08:00
Brennan Bensonandmedox cc1d2e2514 fix(chat): apply window material to Chat UI backgrounds (#40535)
* fix(chat): apply window material to chat surfaces

Use the existing backdrop path for chat canvases and split groups, keeping neighboring terminal backgrounds and readable chat surfaces intact.

Co-authored-by: medox <mohammed.elabidi123@gmail.com>

* test(chat): verify painted backdrop surfaces and isolate fixture state

* Update backdrop imports after renderer module relocation

---------

Co-authored-by: medox <mohammed.elabidi123@gmail.com>
2026-11-10 15:24:54 -08:00
OrcaWinandOrcaWin efc62d1ae5 test: retire low-value creation, title and store cases (#40903)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-10 15:17:39 -08:00
Kelvin Amoaba afa538b3ab fix(codex): stop native chat running Orca's no-op Codex status hooks (#40904)
Native chat strips pane env, so Orca's Codex hooks only spawn a shell per
event. On Windows with Store PowerShell first on PATH, Codex fails every
command-hook spawn with os error 5 (openai/codex#47810), which native chat
showed after each tool call. Disable just Orca's hooks.json entries for the
app-server launch with a session -c hooks.state override; user hooks keep
running.
2026-11-10 23:13:06 +00:00
Neil 633e0727ce refactor(renderer): group workspace and feature module owners (#40895)
* refactor(renderer): group workspace and feature module owners

* refactor(renderer): update retained gate command to moved owner
2026-11-10 15:11:31 -08:00
Kelvin Amoaba 19d0f15919 fix(native-chat): show Codex's model list when the Codex command points at a file in the project (#40897)
* fix(native-chat): list Codex models in the chat's folder when its command line needs that folder

A Codex command such as `node node_modules/@openai/codex/bin/codex.js` failed from Orca's empty probe folder, so the picker fell back to the built-in list. Like Claude's probe, a command line that may name files relative to the chat's folder now lists there, with that folder untrusted so its project config never changes the account's list.

* fix(native-chat): untrust the lowercased folder key Windows Codex looks up first
2026-11-10 23:03:55 +00:00
Neil bda5b082db relay: the edge evacuation sweep leaves region scope to the assign path (#40898)
The assign path judges edge cohort membership with the region the desktop asks for; the sweep used
the stored region preference, which can miss or lag it. A host whose two disagreed was placed on the
edge, drained by the sweep, and placed back on its next ask: at 1% this drained 12 of ~31 moved
hosts within 10 minutes, and each director would repeat it every cooldown. The sweep now ignores
region scope (it still drains hosts that left the cohort by mode, deny or share); a host that asks
for a region outside the scope leaves on that ask.
2026-11-10 15:01:40 -08:00
Kelvin AmoabaandMazurov Anton 0da33cc4fb fix(acp): retry OpenCode's model list when session/new fails while providers are opening (#40894)
* fix(acp): retry transient -32603 during session model catalog probes

opencode answers ACP session/new with a transient -32603 while its
providers still open, and a single probe failure degraded the chat model
picker to one pinned model. Retry initialize and session/new up to
3 times with abort-aware backoff, mirroring the existing sweep retry.

* fix(acp): address CodeRabbit review on transient retry

Re-check operation currency after the backoff sleep so a cancelled or
rotated probe never issues another request, cap retries at three
attempts total (two backoffs), document the helpers, and cover the
abort-during-backoff case with a test.

* fix(acp): retry only OpenCode's session/new and use the abortable timer

* test(acp): list OpenCode's catalog after a -32603 session/new

---------

Co-authored-by: Mazurov Anton <anton.mazurov@reksoft.com>
2026-11-10 22:55:18 +00:00
Neil 21bdae16ab relay: renew the relay token right after wake when it fell due during sleep (#40885)
* relay: renew the relay token right after wake when it fell due during sleep

Node timers stop while the machine sleeps, so a renewal due during a sleep fired that much later
after wake, after the token's wall-clock expiry. With the host-sent keepalive the control survives
the sleep, so the relay closed the live control 4401 a few minutes after wake (cutting any phone
session started since). The renewal and its retries now use a wall-clock timer that runs a due
task on system resume.

* relay: re-arm the renewal from the wall clock on a resume before it is due

A sleep that ends before the renewal is due still cost the timer clock the whole sleep, so the
renewal fired that much late. On resume a task not yet due is re-armed for the wall-clock time left.
2026-11-10 14:54:48 -08:00
OrcaWinandOrcaWin 7a09f21423 test: remove low-value terminal and store examples (#40892)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-10 14:53:09 -08:00
Kelvin Amoaba e844a85e8a fix(native-chat): keep Codex generated images visible when a turn folds (#40893)
The settled-turn fold kept only the last prose row, so a completed Codex image generation folded behind the final answer. Mark generated-image rows and keep them out of the fold, like plans and turn notices.
2026-11-10 22:52:41 +00:00
Kelvin Amoaba ddab3a6e1a fix(codex): deliver handoffs when Codex's status line paints nothing (#40888)
* fix(codex): take Codex's idle report as its live composer so handoffs reach a blank status line

* fix(codex): drop the idle verdict when Codex reports blocked, error or clear
2026-11-10 22:47:11 +00:00
Jinwoo Hong 144e0e4ce2 fix(cloud): send the refresh token as proof when creating a linked cloud profile (#40713)
The profile request now carries the caller's refresh token alongside the
access token, so the server can require proof of the long-lived sign-in
before issuing the new profile's session. Servers that do not expect the
field ignore it.
2026-11-10 17:46:44 -05:00
Jinwoo Hong 0a989ce9d2 fix(cloud-auth): keep users signed in through transient auth errors (#40837)
* fix(cloud-auth): keep users signed in through transient auth errors

Only the auth server's JSON 401 invalid_refresh_token verdict now ends a
desktop sign-in. 403s, HTML or other non-JSON answers, other 401 codes,
5xx, 429, timeouts and network errors keep the session and retry later.

An unanswered refresh is now retried once within the server's rotation
grace (20s timeout, 10s pause, one retry) instead of replaying the token
90s later, when the server treats it as reuse.

A wrapped request that still answers 401 right after a successful refresh
returns its error instead of discarding the session the server just
accepted.

* fix(cloud-auth): refresh only on a 401, at most once per backoff window

A 403 from a signed request is a permission answer, so it no longer
spends a refresh-token rotation. A request that keeps answering 401
forces at most one refresh per session every 5 minutes; concurrent
rejections join the refresh already in flight or reuse a newer token,
and later ones return their 401.
2026-11-10 17:46:40 -05:00
Kelvin AmoabaandYangNuoCheng c2e6ebbf0b feat(cli): read, message and wait on agent chats with orca chat list/read/send/wait (#40836)
* feat(cli): read and message agent chats with orca chat list/read/send

Carried over from #33115 (squashed onto current main; conflicts with chat rename resolved).

* feat(cli): wait for a chat to settle, read text from stdin, refuse paired hosts

chat send --wait and chat wait poll the history's working flag until the
chat is idle or waits on a question, keeping a queued message pending until
it runs. chat read is one history fetch and returns a cursor. Unattached
and paired-host refusals say what they mean.

* fix(cli): settle a held queued message, skip subagent rows, resolve chat list --worktree

* fix(cli): show chat errors, keep waiting through queue handoffs, return the whole settled turn

* fix(cli): settle any queued message the host will not send, show grouped questions and only answerable prompts

* test(cli): drive chat tests through a real client and fake timers

* fix(cli): describe chat --session, --wait and --state in command help

* fix(cli): report chat wait in chat list's states, reject unknown --state, flag truncated reads

* docs(cli): say chat wait settles on the chat's own turn; no truncated flag on held

* fix(cli): chat rename help asks for a session id, not a handle

---------

Co-authored-by: YangNuoCheng <42427439+YangNuoCheng@users.noreply.github.com>
2026-11-10 22:37:31 +00:00
Neil b5611915f9 refactor(mobile): group usage and shell generation owners (#40783)
* refactor(mobile): group usage and shell generation owners

* refactor(mobile): use direct module types in reset credit adapter
2026-11-10 14:28:58 -08:00
OrcaWinandOrcaWin 60d7605df4 Remove low-value test examples and two suites (#40883)
Co-authored-by: OrcaWin <293788423+OrcaWin@users.noreply.github.com>
2026-11-10 14:20:52 -08:00
Kelvin Amoaba e319da9cef fix(mobile): save the downloaded desktop page on Android 7 (#40878)
* fix(mobile): save the downloaded desktop page on Android 7

Below API 26, expo-file-system 55 moves a directory with File.copyTo, which
creates an empty directory, so committing a downloaded page build always
failed with cache-write-failed. Copy the tree recursively and delete the
source there instead.

* fix(mobile): remove a partial copy when an Android 7 page move fails
2026-11-10 22:13:18 +00:00
Neil 557d588e69 relay edge: tag auth-expiry closes and control opens with keepalive use and host silence (#40882)
Desktop closes for expired relay authorization rose after the keepalive release (3/day to ~35/day
on the edge cohort). The close metric now carries whether the control took the keepalive offer and
how long the host had been silent, and control_active carries keepalive use, so the cause can be
told apart (stalled desktop timers vs a control whose refresh stopped while it stayed live).
2026-11-10 14:08:25 -08:00
Kelvin Amoaba 6dfb4dc60d feat(cli): add orca tab annotations so agents can read a tab's unsent notes (#40873)
Prints the notes the user left on a browser tab and has not sent, the same text the tab's Copy button gives (--json returns them). Reads the saved session, so it works for closed-and-restored tabs; SSH hosts reach it through the desktop browser bridge. A page drawn by a paired desktop reports that its notes stay there.
2026-11-10 22:02:43 +00:00