After Cmd+Q quit + reopen, sessions the user killed reappeared as alive.
The session_worktrees "shutdown-saved" restore marker was write-only:
RestoreAll auto-relaunches any terminated session that still has a marker,
but Kill never deleted it and RestoreAll never consumed it. A session that
survived one reopen cycle (gaining a marker via reconcileLive) and was then
killed still carried the marker, so the next boot resurrected it.
Two-layer fix:
- Kill now deletes the marker (best-effort) after MarkTerminated: a user
kill is explicit terminal intent and must not leave a resurrection marker.
- RestoreAll deletes the marker after consuming preserveRef and attempting
relaunch, making it one-shot so it never outlives a single restart. A
still-live session re-acquires a fresh marker at the next quit.
Manual Restore stays marker-independent, so killed sessions remain
restorable on demand, just not auto-resurrected on boot.
Adds DeleteSessionWorktrees to the Manager Store interface (the concrete
store already implements it) and tests covering all three behaviors.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* feat(spawn): add required --name flag for sidebar display name
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(spawn): require --name for sidebar label; cap at 20 chars
Add a required --name flag to `ao spawn` that sets the session's
sidebar display name. The CLI rejects a missing or >20-character name
before contacting the daemon.
The daemon's POST /sessions keeps displayName optional (the desktop
new-task dialog omits it and the read model falls back to the session
id) but enforces the same 20-character cap when present, so a direct
API call cannot exceed it. The value flows CLI -> SpawnSessionRequest
-> SpawnConfig -> session record, and the existing read-model fallback
(displayName ?? issueId ?? id) renders it in the sidebar unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Two PRs (#2193, #2200) merged ~2h apart each added a migration
file numbered 0020, since each branch only saw one such file at
CI time and neither was rebased against the other before merge.
Renumbers 0020_pr_reviews.sql to 0021_pr_reviews.sql.
TestMigrationVersionsAreUnique already existed and is correct;
it didn't catch this because it only runs against each PR's
branch state, not the merged result of both PRs together.
The reviewer's read-only guarantee was enforced only by the prompt. Add
AllowedTools/DisallowedTools to ports.LaunchConfig and plumb them through
the claude-code agent adapter to --allowedTools/--disallowedTools (each list
comma-joined into one value so a rule containing spaces like "Bash(git
diff:*)" is not split into separate tool names). Empty lists emit nothing, so
worker sessions are unaffected.
Launch the reviewer off bypassPermissions (which skips the permission system
and ignores allow/deny rules) in the default auto mode, with an allowlist
scoped to Read/Grep/Glob and the few Bash commands a reviewer needs (gh, git
diff/log/show/status, ao review submit) and an explicit deny list for
Edit/Write/NotebookEdit/git push/git commit as defense in depth.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* fix(ui): use popper position for SelectContent to prevent scroll reset
* fix(ui): use React.memo to prevent scroll resets without breaking popper position
* fix(ci): pin publish step to bash so the retry loop runs on Windows
The 3x retry wrapper added in #2266 is bash syntax, but the release
matrix Publish step inherited the runner default shell, which is
PowerShell on windows-latest. That made every Windows publish fail with
a ParserError. Pin shell: bash on all four publish steps.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(ci): point fresh-install smoke check at an empty release repo (#2267)
The container check asserts `ao start` fails cleanly on a fresh box with
no published asset. Now that AgentWrapper publishes a linux-x64 AppImage,
an unpinned smoke binary downloads it and exits 0, tripping the
assertion. Build the smoke binary against a release repo with no assets
so the fetch path deterministically 404s and start exits non-zero with a
clear error, preserving the test's intent without depending on what the
real repo publishes.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Releases fail intermittently on transient macOS issues that hit EITHER
macOS leg, not just Intel: Apple notary -1009 connection-offline (seen on
macos-latest/arm64) and osx-sign keychain races, code object is not signed
at all (seen on macos-15-intel). Both pass on a manual re-run.
Wrap npm run publish in a 3x retry with backoff in all four release/nightly
publish steps so these self-heal in-place instead of failing the run (and,
for nightly, skipping the whole feed). Keeps the Intel leg coupled so the
x64 feed entry stays guaranteed.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Nightly builds publish as GitHub prereleases. configureFeed set channel
and allowDowngrade but never allowPrerelease, which defaults to false, so
electron-updater only inspected the latest NON-prerelease release
(e.g. v0.10.0) and 404d looking for nightly-mac.yml there. The nightly
channel therefore never resolved, regardless of whether a nightly was
published.
Set allowPrerelease = (channel === nightly): nightly scans prereleases
and finds the nightly feed; stable stays on non-prereleases only.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* fix(update): generate app-update.yml at build time; drop runtime setFeedURL
electron-forge does not emit app-update.yml; electron-updater requires it at
process.resourcesPath to resolve the GitHub release feed. Without it every
packaged app threw ENOENT on the first download attempt, making updates
detected but never installed.
Two-part fix:
- forge.config.ts: add postPackage hook that writes app-update.yml into
each platform's Resources dir (baked from AO_RELEASE_REPO so fork builds
point at the fork, prod at AgentWrapper/agent-orchestrator).
- auto-updater.ts: remove setFeedURL + repo() + DEFAULT_RELEASE_REPO;
configureFeed now only sets channel + allowDowngrade. electron-updater
auto-loads the bundled yml on the first checkForUpdates call.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
* fix(update): generate app-update.yml before signing, not after
The postPackage hook wrote app-update.yml into Contents/Resources AFTER osxSign
sealed the bundle, so the added file was unsealed and macOS reported the app as
"damaged" (codesign: "a sealed resource is missing or invalid"). Generate it in
a prePackage hook and ship it via packagerConfig.extraResource, so it is copied
into the bundle and signed as part of the seal. The generated app-update.yml is
gitignored.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
* ci(release): move Intel leg to macos-15-intel and gate the feed on it
macos-13 is deprecated and has no runner capacity (multi-hour queues),
so the detached release-intel leg never completed. Switch it to the
supported macos-15-intel image in both the release and nightly workflows.
Now that the Intel leg gets a runner and builds + signs the x64 installer
reliably (verified on fork v0.10.10), re-couple it into publish-feed
(needs: [release, release-intel]) so latest-mac.yml / nightly-mac.yml
always carry the x64 entry and Intel macOS users receive auto-updates.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Stable release builds read the version from frontend/package.json (the
release workflow does not stamp from the git tag, unlike nightly), so the
0.10.0 prod cut needs this committed at the tagged commit. Otherwise the
app ships as 0.0.0 and the stable update channel never advances.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
The Desktop release fires on any desktop-v* tag push and consumes the
repo-level signing secrets, so any write-access collaborator can cut a
signed prod build. Reference the `release` environment from the build
jobs so a designated reviewer must approve before the secrets are used.
Takes effect once required reviewers are configured on that environment.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Remove macos-13 from the release matrix in both frontend-release.yml and
frontend-nightly.yml, replacing it with a standalone release-intel job
that runs on macos-13 independently. publish-feed keeps needs: release
(the three fast legs), so it no longer waits hours for the scarce Intel
runner. The darwin-x64 installer and stable alias are still built and
uploaded; feed.mjs includes the x64 entry on a best-effort basis.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* feat(settings): IPC bridge for global update settings (get/set)
Expose readUpdateSettings/writeUpdateSettings to the renderer via
updateSettings:get/set so the Global Settings page can read and persist
the auto-update channel choice. Refs #2207.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(cli): show ao start download + install progress
The fetch path streamed a hundreds-of-MB asset and ran a silent install
with zero output, so a slow download looked like a hang. Print a start
line (asset, ~size, repo), stream a TTY-aware progress reader (live \r
percentage on a terminal, start+done lines off a TTY), and announce the
otherwise-silent unpack/install step. stderr only, so --json stdout stays
clean; no new dependency. Closes#2234.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(settings): Updates (channel) + Migration sections in Global Settings
- UpdatesSection: pick Stable vs Nightly auto-update channel, persisted to
~/.ao/update-settings.json via the new IPC bridge. Closes#2207.
- MigrationSection: drop-in card showing migration status + last report/error
and a Run/Re-run button hitting the idempotent POST /api/v1/import; works
even after Don't Migrate. Closes#2205.
- Wire both into the (previously blank) GlobalSettingsForm; cover both with tests.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
* fix(browser): re-measure preview bounds after layout transitions
The native WebContentsView preview is positioned from a renderer-measured
rect, re-measured on ResizeObserver + window resize/scroll. But a
ResizeObserver fires on size changes only, so a position-only layout shift
(entering/leaving pop-out moves the slot to another panel; opening the
inspector, which ao preview does, reflows its x) left the overlay at stale
bounds, visibly spilling over the sidebar/terminal until an unrelated resize
fixed it. Drive a settle re-measure on mount and on active/pop-out
transitions (immediate frame + once the ~240ms panel transition settles) so
the final geometry always wins.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(backend): gofmt + goimports pre-existing files (fix red CI)
config.go, project_test.go, importer_test.go and spawn_windows.go were
committed unformatted earlier and fail go.yml's gofmt check and golangci's
goimports formatter on main. Apply `golangci-lint fmt`; formatting only, no
logic change. Unblocks this PR's lint/build-test jobs.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(settings): manual Check for updates + Update button
Auto-update only checked at launch, so users with auto-updates off (or who
just want it now) had no way to update from the app. Add an on-demand flow:
checkForUpdatesNow/downloadUpdateNow/quitAndInstallUpdate in auto-updater
(works regardless of the opt-in; not-packaged dev surfaces 'unsupported'),
forwarded to the renderer as a live UpdateStatus over updates:status. The
Updates section gains a Check for updates button and an Update button that
downloads then installs (Restart & install), with checking/available/
downloading%/downloaded/error status. Refs #2207.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
* fix(sidebar): always offer Global settings in the footer menu
The footer Settings menu showed either Project settings (with a project
selected) or Global settings (with none), so global settings was
unreachable while inside a project. Always list Global settings; add
Project settings above it when a project is active. Refs #2205.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(settings): show current app version in Updates section
Surface app.getVersion() as a 'Current version vX.Y.Z' line above the
update controls so users can see what they're running. Refs #2207.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
The gitleaks job scans full git history on push to main (gitleaks-action
v1.6.0 runs `gitleaks --path` with no commit filter on push events). The
ReverbCode rewrite graft (#2166) carried 279 commits of prior history, 10 of
which trip 42 findings: example secret-patterns in security docs and
secret-redaction test fixtures. All offending files are already deleted from
the tree, so PRs pass (they scan only their own diff) but every push to main
re-scans history and fails indefinitely.
Add `.github/.gitleaks.toml` (the action's default config-path) carrying the
full gitleaks v7.4.0 default ruleset plus a commit allowlist for the 10
historical commits. Detection on new code is unchanged; only these specific
historical commits are skipped.
Verified with gitleaks v7.4.0: full-history scan goes 42 leaks -> 0.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
waitForStopped removed the run-file and then additionally waited for the
daemon process to fully exit, erroring with "removed run-file but did not
exit within 10s" if it lingered past the stop timeout. The run-file is the
daemon's own liveness marker: once it is gone the daemon has committed to
stopping. When no desktop client is connected the daemon can drain its
background workers slower than the stop timeout, which made ao stop
spuriously report failure (the TestE2E_Lifecycle failure in #2214).
Treat run-file removal as stopped: keep polling for full process exit as a
best effort (so Windows releases inherited handles before callers clean the
data dir) but no longer error when that grace elapses.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Add a verbatim Apache License 2.0 LICENSE file at the repo root, with the
appendix copyright line filled as "Copyright 2026 Untrivial".
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* docs(spec): feed-publishing + macOS signing design (#2220)
Design for the electron-updater feed-publishing workstream plus the macOS
code-signing half of Track B, shipped as one PR so a single review unblocks
auto-update. Sidecar-only post-matrix join job emits latest*.yml / nightly*.yml
+ .blockmap sidecars on all three platforms (no maker changes, no artifact
mutation). macOS signing reproduces the proven local runbook in CI: keychain
provisioning, hardened-runtime entitlements, and notarization via the App Store
Connect API key (.p8) path, with the osxNotarize cast fixed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(plan): feed-publishing + macOS signing implementation plan (#2220)
Six TDD tasks: blockmap wrapper, feed module (selection + yml), osxNotarize
API-key rewire, macOS signing-setup composite action, and the two workflow
wirings (latest + nightly feed jobs). Two evidence-based simplifications vs the
spec: no custom entitlements plist (default osxSign entitlements are proven by
the local runbook) and no Node bump (CI already below the crash ceiling).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(feed): blockmap sidecar wrapper over app-builder-lib
Adds frontend/scripts/blockmap.mjs isolating the single fragile internal
import (app-builder-lib/out/targets/blockmap/blockmap.js) behind a thin
writeBlockmap(filePath) wrapper. Returns { sha512, size } only, omitting
blockMapSize to force the sidecar differential path in electron-updater.
Test added with // @vitest-environment node directive (required: project
vitest config defaults to jsdom; same pattern as nightly-version.test.mjs).
TDD: red on missing module, green after implementation.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(feed): installer selection + electron-updater yml assembly
Implement selectInstallers, feedFilename, and buildYml pure functions
for packaging versioned installers and generating platform-specific
feed metadata. Excludes ao-start aliases and deb/rpm packages.
Generates blockmap sidecars and yml with deprecated top-level
path/sha512 for compatibility, no blockMapSize (forces sidecar diff).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(sign): notarize via App Store Connect API key; drop osxNotarize cast
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(ci): macOS signing-setup composite action
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(ci): sign macOS + publish latest feed in the release workflow
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(ci): sign macOS + publish nightly feed in the nightly workflow
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* docs(spec): auto-update with stable + nightly channels
Design spec for channel-aware auto-update via electron-updater: stable
(manual semver releases) and nightly (daily CI cron, version
X.Y.(Z+1)-nightly.<UTC-ts>+<sha>). Covers runtime wiring, the feed,
the version scheme and its ordering/channel rationale, the nightly
pipeline, in-app UX (opt-in + channel + nightly disclaimer), error
handling, testing, and the deferred/prereq items (CI signing for macOS
updates, stable version stamping, polished UI in #2207).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(plan): auto-update channels implementation plan
Bite-sized TDD plan for the in-app channel-aware updater + nightly pipeline:
nightly version compute module, daily cron workflow, macOS feed metadata,
~/.ao update-settings module, electron-updater shell + main.ts wiring, and the
first-run opt-in/channel/nightly-disclaimer prompt. Deferred items and Track-B
prereqs (CI signing, stable stamping) are called out.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(release): nightly version compute module
computeNightlyVersion -> X.Y.(Z+1)-nightly.<UTC-ts>+<sha>: next-patch base,
fixed-width UTC timestamp for monotonic prerelease ordering, sha as build
metadata. Pure ESM so CI runs it and vitest tests it.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(release): daily nightly build + publish workflow
Cron computes the nightly version from the latest stable tag, stamps it,
builds all platforms, and publishes a prerelease (nightly channel). Skips
when HEAD is already covered by the latest nightly. forge publisher prerelease
flag is now env-driven (AO_RELEASE_PRERELEASE) so stable stays non-prerelease.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(release): pass nightly version via env, not inline interpolation
Avoids interpolating ${{ }} into the run script (GitHub Actions injection
hardening); the version now arrives via the NIGHTLY_VERSION env var.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(update): persist auto-update settings under ~/.ao
readUpdateSettings/writeUpdateSettings store {enabled, channel, nightlyAck}
with safe defaults and channel coercion, atomic temp+rename like app-state.ts.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(update): channel-aware electron-updater wiring
Replace update-electron-app (stable-only, no channels) with electron-updater
driven by the user's ~/.ao settings: channel from settings, allowDowngrade for
channel switches, auto-download gated on opt-in, errors swallowed. Feed
configured via setFeedURL since forge does not emit app-update.yml.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* style(update): tabs in auto-updater.ts to match repo prettier config
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(update): first-run opt-in + channel + nightly disclaimer prompt
Minimal main-process dialog flow: opt into auto-updates, pick stable/nightly,
and acknowledge a nightly instability/data-loss disclaimer. Persists the choice
to ~/.ao. Polished Settings-page selector is tracked in #2207.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(spec): correct feed-yml premise (no platform emits it as built)
Final review verified that publish:null in both custom makers makes
electron-builder skip update-info yml generation on Windows and Linux too
(app-builder-lib PublishManager returns early before createUpdateInfoTasks),
not just signing-blocked macOS. The runtime updater is therefore inert on all
platforms until a separate feed-publishing workstream (yml gen+upload across
all 3 platforms, coupled to Track-B signing) lands. Recorded in the Feed
section and the prerequisites table.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(release): nightly skip-guard must match the v<version> tag namespace
The forge publisher tags nightly releases v<version> (e.g.
v0.10.4-nightly.<ts>), not desktop-v..., so the no-new-commits guard queried a
namespace that never exists and never fired. Query v*-nightly.* instead.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
* chore(release): run nightly at 19:00 IST (13:30 UTC)
Was 03:00 UTC (08:30 AM IST); move to 13:30 UTC = 7:00 PM IST.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* docs: design for dashboard legacy-migration popup + app-state marker
Spec for the app-side migration trigger (Approach A): projects-only import
daemon API + a migration marker in ~/.ao/app-state.json, with a launch-time
popup (Proceed / Skip / Don't Migrate). Settings redo path deferred to #2205.
Includes the projects-only import-offer backend plan it consumes.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs: implementation plan for legacy-migration popup + marker
Part A reuses the projects-only import API (import-offer plan) with an
availability-only Status; Part B adds the app-state migration marker (schema v2),
IPC, the useMigrationOffer gate, and the MigrationPopup (Proceed/Skip/Don't
Migrate). Settings redo path deferred to #2205.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* refactor(legacyimport): scope import to projects + settings only
Remove orchestrator/transcript import code (orchestrator.go, claude.go,
session_import_store.go and their tests). Trim Store, Options, Report to
projects-only fields. Drop defaultClaudeProjectsDir and projectSessionsDir
from paths.go. Add yaml.TypeError robustness in config.go. Update cli/import.go
confirm prompt and summary. Update importer_test.go to projects-only assertions.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(importer): availability probe + projects-only run
Create service/importer.Manager with Status (physical availability check only,
no DB heuristic) and Run (delegates to legacyimport.Run). The app-state.json
marker governs whether to prompt; this service only answers whether legacy data
is physically present.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(httpd): import controller + DTOs (GET/POST /api/v1/import)
Add ImportStatusResponse/ImportRunResponse DTOs to dto.go. Create
ImportController with GET (status probe) and POST (run) handlers, both
returning 501 when Svc is nil. Wire APIDeps.Import + API.imports in api.go.
Add controller tests (status, status error, run, run error).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(apispec): describe /api/v1/import; regenerate openapi + schema.ts
Add import tag, importOperations() (GET + POST /api/v1/import), and schema
name mappings (ImportStatusResponse, ImportRunResponse, ImportReport) to
build.go. Regenerate openapi.yaml and frontend/src/api/schema.ts. Route-spec
parity test passes. Restore the nil-svc-501 import controller test now that
the spec includes the operation.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(daemon): mount import service on the API
Wire importsvc.New(importsvc.Deps{Store: store}) into APIDeps.Import in
daemon.go so the daemon serves GET/POST /api/v1/import backed by the live
sqlite store. Projects-only; no DataDir.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(cli): drop resolved §6.4 first-boot-import TODO
The legacy import is now a daemon API (GET/POST /api/v1/import) served by
the importer service and the desktop app handles the popup prompt via the
app-state.json migration marker. The TODO comment is resolved.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(legacyimport): remove dead isDir helper
isDir was only used by the deleted projectSessionsDir function.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: update package-lock.json after npm install for api:ts
openapi-typescript was missing from root node_modules; npm install
populated it so npm run api:ts could regenerate schema.ts.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(httpd): drop em-dash comment + unused ImportStatusResult alias
Review findings M1/M2 from G1 task review.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(app-state): migration marker (schema v2) + updateMigration
Bump SCHEMA_VERSION to 2, add MigrationStatus/MigrationState types and
migration? field to AppStateMarker. Extract atomicWriteMarker helper and
reuse it. writeAppStateMarker now preserves an existing migration block
across launch writes. Add updateMigration (IPC setter) and readMigrationState
(IPC getter) exports. TDD: tests added first (red), then implementation (green).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(ipc): expose app-state migration getter/setter to the renderer
Add appState:getMigration / appState:setMigration IPC handlers in main.ts.
Add ao.appState.getMigration / setMigration to preload.ts (typed via AoBridge).
Add appState preview fallback in bridge.ts and test setup so AoBridge stays
satisfied in both browser preview and test environments.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(app-state): cover corrupt-marker case + clean up temp dirs
G2 review findings I2 (corrupt-JSON branch untested) and m1 (temp dirs not cleaned).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(renderer): useMigrationOffer gate (marker + availability)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(renderer): MigrationPopup (Proceed / Skip / Don't Migrate)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(renderer): surface MigrationPopup on the dashboard
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
The desktop release only built on macos-latest (Apple Silicon, arm64),
so the darwin-x64 stable alias was never produced and Intel-Mac
ao start 404s fetching agent-orchestrator-darwin-x64.zip.
Add the macos-13 (Intel x64) runner to the job matrix and broaden the
macOS alias-upload step's condition from matrix.os == 'macos-latest' to
startsWith(matrix.os, 'macos') so it runs on both macOS runners. The
step already derives arch via uname -m, so the Intel runner now emits
and uploads the x64 alias. Updated the now-stale gap comment.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* feat(frontend): add blank Global Settings page
Add a Global Settings page to the desktop renderer, reachable from the
sidebar Settings menu when no project is selected. The project-scoped
Settings entry stays when a project is active; otherwise the menu shows
Global settings, which opens the new /settings route. The page body is
intentionally empty for now and will be filled in incrementally.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(frontend): regenerate route tree + test Global Settings menu
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* feat(start): implement ao start fetch/open for Windows and Linux
Fill in the non-darwin branches of the bootstrapper (T6/T7):
- assetName() selects the per-GOOS stable release asset: windows ->
agent-orchestrator-win32-x64.exe (NSIS installer), linux ->
agent-orchestrator-linux-x64.AppImage. amd64-only for now via
requireAMD64(), which returns a clear unsupported-arch error.
- fetchApp() dispatches per GOOS. Windows downloads the NSIS installer and
runs it silently (/S) to the default per-user dir, then resolves the
installed exe under %LOCALAPPDATA%\Programs. Linux downloads the AppImage
to a stable path under ~/.ao (atomic temp+rename), chmods it executable,
and skips any install step so re-runs resolve without re-fetching.
- knownAppLocations() scans the per-user and per-machine Windows install
dirs and the stable Linux AppImage path.
- isUsableBundle() treats a win exe / linux AppImage as a regular file
(darwin stays a directory).
- openApp() launches win/linux detached via the existing StartProcess seam,
forwarding --installed-via=npm-bootstrap, and falls back to manual-open on
spawn failure.
The Windows silent-install path is marked ponytail (untestable on the macOS
build host); a wrong install dir surfaces as a clear not-found error. Tests
cover asset naming, arch gating, scan locations, regular-file vs dir, and
the detached-spawn + fallback paths.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(release): build + publish a Linux AppImage for ao start (T7)
AppImage is the Linux fetch-and-run artifact for the bootstrapper (spec
§11.3): a single self-contained executable ao start downloads and runs
directly, no system package manager.
- makers/maker-appimage.ts: a MakerBase subclass bridging to
electron-builder's buildForge (appImage target), mirroring maker-nsis.ts,
since Forge has no first-party AppImage maker. publish:null so Forge owns
release uploads.
- forge.config.ts: register MakerAppImage for linux; keep deb/rpm for users
who want a system package.
- frontend-release.yml: on ubuntu-latest, copy the built AppImage to the
stable, space-free name agent-orchestrator-linux-x64.AppImage and upload
it to the v<version> release with --clobber, mirroring the Windows step.
Build-untested on this macOS host: the first ubuntu CI run must confirm the
electron-builder AppImage target token and the out/make/*.AppImage output
glob.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(ci): green up ao start Win/Linux PR (lint, cross-OS test, container smoke)
Five golangci-lint findings in start.go, two cross-OS test failures, and the
fresh-install container smoke were broken by the Win/Linux bootstrapper diff.
Go lint (start.go):
- gocritic filepathJoin: build "/Applications/<bundle>" as a literal instead of
filepath.Join with a separator-bearing arg.
- gocritic httpNoBody: pass http.NoBody, not nil, to NewRequestWithContext.
- gosec G302: annotate the AppImage chmod 0755 with a nolint; an AppImage is a
self-contained executable and must be executable.
- nilerr: annotate openApp's intentional (false, nil) on launch failure; the
failure is reported via the bool, not as an error.
- unparam: resolveApp's error result was always nil; drop it and update callers.
Cross-OS tests (start_test.go):
- makeBundle created a directory, which only stats as usable on macOS. Make it a
regular file on Windows/Linux so the marker/scan resolve tests pass there,
matching isUsableBundle's per-OS rule.
Container fresh-install smoke (test/cli/install-check.sh, Dockerfile):
- ao start is now the desktop-app launcher and no longer runs a daemon, so the
old daemon status/shutdown/stop assertions could never pass. Assert instead
that on a fresh box start reaches the fetch path and exits non-zero with a
clear error (404 download on amd64, unsupported-arch on arm64). Refresh the
stale daemon-reaping comments in the Dockerfile.
Verified locally: go build/vet ok, golangci-lint v2.12.2 reports 0 issues,
go test -tags e2e ./internal/cli/... passes (the only remaining failure,
TestE2E_Lifecycle, is a pre-existing daemon-shutdown flake that fails the same
way on upstream/main on this host), and the container smoke passes on both
linux/arm64 and linux/amd64.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* docs: grounded spec for ao start bootstrapper + npm deprecation
Real-codebase-verified implementation spec (Track A: launcher + app-state
marker + release asset wiring). Replaces the somthing.md draft's aspirational
assumptions with file:line ground truth: correct bundle name (Agent
Orchestrator.app), publish repo (aoagents/agent-orchestrator), draft-release
+ asset-rename gaps, unsigned-build reality, and the already-wired
update-electron-app updater. Includes a dependency-ordered task breakdown
for AO execution.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs: reframe spec around Go ao start subcommand (not a JS launcher)
The npm package ships the existing Go cobra CLI (backend/cmd/ao); this effort
rewrites the `ao start` subcommand to fetch+open the desktop app. Corrections:
- releases land on AgentWrapper/agent-orchestrator (aoagents was the temporary
rewrite home; forge publisher must be repointed)
- ao start stops starting the daemon; the frontend owns the daemon
- adds the real Go CLI command surface (1.7), the npm-delivery gap for the Go
binary (1.6), and the legacy first-boot import decision (6.4)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs: separate prod vs test release targets in ao start spec
Dev/test loop must never cut a production release. Download repo + forge
publisher + npm scope are now build-time overridable:
- prod: AgentWrapper/agent-orchestrator + real package name
- test: harshitsinghbhandari/agent-orchestrator (fork) + @theharshitsingh/ao
T3/T5 now release+test against the fork and the test scope, with prod cut as a
separate gated step.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(cli): rewrite ao start to fetch + open the desktop app (T1)
ao start no longer starts the daemon (the desktop app owns it). It now
resolves the installed app (~/.ao/app-state.json marker -> stat -> known-
location scan), fetches the latest release zip and ditto-unpacks it on macOS
when absent, opens it with --installed-via=npm-bootstrap, and prints an honest
deprecation notice. releaseRepo is build-time overridable (-ldflags) so test
builds fetch from the fork. Windows/Linux fetch/open are stubbed for T6/T7.
Review fixes folded in: download() copies deps.HTTPClient and drops its 2s
loopback-probe timeout (a real release asset is hundreds of MB), and fetchApp
clears ~/.ao/staging before unpacking. Regression test covers the timeout.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(desktop): write ~/.ao/app-state.json marker on launch (T4)
The app is the sole writer of the marker ao start reads to locate the bundle.
New frontend/src/main/app-state.ts does an atomic temp+rename write mirroring
the daemon's runfile.Write, preserving installedAt/installSource across launches
and refreshing appPath/version/lastReconciledAt. main.ts hooks it into
app.whenReady ordered relocate (macOS) -> write marker -> createWindow, both
non-fatal. Bundle path is derived from process.execPath (not app.getAppPath,
which is the asar path); JSON keys match start.go's appState reader exactly.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* ci(desktop-release): overridable repo, published release, stable asset aliases (T3)
Make the bootstrapper's constant releases/latest/download URL resolve:
- forge publisher repo reads AO_RELEASE_REPO (default AgentWrapper/agent-
orchestrator); the workflow sets it to github.repository so a fork run
publishes to the fork and never to prod
- draft:false so the release is immediately live (constant URL needs it)
- add ubuntu-latest to the matrix (issue #2191)
- post-publish steps upload stable space-free aliases
(agent-orchestrator-darwin-arm64.zip, agent-orchestrator-win32-x64.exe)
matching exactly what start.go fetches
Review fix: the alias upload targeted GITHUB_REF_NAME (the git tag), but
publisher-github creates the release as v<package.json version>; retargeted to
that and relaxed the guard so workflow_dispatch also produces aliases.
Known gaps (documented inline): macOS x64 needs an Intel runner (macos-latest is
arm64-only); the Linux stable asset name awaits the deb/rpm-vs-AppImage decision.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
* feat(npm): deliver Go ao binary as @aoagents/ao via per-platform packages (T2)
Ship the ao CLI to npm with ZERO install scripts using the esbuild model:
- packages/ao: pure-JS @aoagents/ao with a bin/ao.js shim that resolves and
execs the matching @aoagents/ao-<platform>-<arch> optionalDependency
- four platform packages (darwin-arm64/x64, win32-x64, linux-x64), each os/cpu
gated so npm installs only the host's; binary cross-compiled CGO-free
(modernc.org/sqlite), gitignored, shipped via files
- build-binaries.sh cross-compiles all four; releaseRepo keeps its prod default
Not added as root workspaces on purpose: os/cpu-restricted members make root
`npm ci` fail EBADPLATFORM (CI's api-drift job runs it). Packages publish
standalone; the shim was verified in a published-like layout.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(npm): set @aoagents/ao packages to 0.10.0 (above existing 0.9.5)
The existing @aoagents/ao on npm is 0.9.5; the launcher must publish a higher
version so npm latest advances and existing users get the new fetch-and-open
binary on update. All five packages plus the four optionalDeps pins set to
0.10.0.
* chore: format with prettier [skip ci]
* fix(start): capture install provenance before macOS relocation
moveToApplicationsFolder() relaunches the app from /Applications without
forwarding the --installed-via arg, and code past a successful move never
runs in the staging instance. The post-move instance therefore wrote
installSource="unknown", and writeAppStateMarker's sticky logic then locked
it there, losing the npm-bootstrap provenance in the exact path it exists for.
Write the marker before relocation when --installed-via is present so the
source is persisted while the arg is still available; the post-move launch
preserves it (sticky installSource) while refreshing appPath to /Applications.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* fix(browser): clip preview view to its panel column so it can't paint out of bounds
The inspector's Browser tab hosts a native WebContentsView positioned from the
slot's getBoundingClientRect. That overlay is a window-level layer, so DOM
overflow:hidden never clips it. The slot sits inside the inspector's
min-w-[280px] reflow guard, so on a narrow panel (small window, or mid
collapse/expand) the slot's box spills past its resizable-panel column and the
view paints over the terminal/sidebar, "out of the box" (#2202).
Two fixes in useBrowserView:
- Intersect the measured slot rect with its [data-panel] column before sending
bounds, so the view can only ever paint inside that column.
- Observe the panel column (not just the slot) with the ResizeObserver. During
the flex-grow collapse/expand animation the slot's width stays pinned by
min-w, so a slot-only observer never re-fires and the bounds go stale; the
column's width changes every frame, so observing it re-measures throughout.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* fix(daemon): do not tear down live sessions on shutdown; adopt them on boot
Remove SaveAndTeardownAll from the graceful shutdown path. Live tmux/ConPTY
sessions survive daemon exit; Reconcile on the next boot adopts them via
reconcileLive, preserving session IDs and preventing the id-increment bug.
Add runShutdownSessionLifecycle as a testable seam and narrow the
sessionLifecycle interface to Reconcile+RestoreAll only.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(daemon): make shutdown-teardown regression test falsifiable
Re-add SaveAndTeardownAll to the sessionLifecycle interface and to
fakeSessionLifecycle so TestShutdown_DoesNotCallSaveAndTeardownAll is
genuinely falsifiable: the flag flips if runShutdownSessionLifecycle
ever calls sl.SaveAndTeardownAll, making the assertion meaningful.
Name the sl parameter (was discarded with _) so the seam is visible.
RED: with a temporary sl.SaveAndTeardownAll call, test fails.
GREEN: without it, test passes. go test ./internal/daemon/... -race: 23/23.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* refactor(daemon): guard shutdown-teardown at compile time via narrowed interface
Remove SaveAndTeardownAll from sessionLifecycle so daemon.Run physically
cannot call teardown on shutdown. Delete the no-op runShutdownSessionLifecycle
seam and its test. The narrowed interface is the guard: re-introducing teardown
requires a visible, reviewable interface change.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(session): make ensure-orchestrator idempotent so POST cannot mint a duplicate
When clean=false, SpawnOrchestrator now checks for an existing active orchestrator
and returns it directly instead of always calling Spawn. Adds RED/GREEN tests.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* test(session): drop redundant NoCleanSkipsKills, covered by SpawnsWhenNoneExists
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(daemon): supervisor watchdog
Adds backend/internal/daemon/supervisor: a transport-agnostic watchdog
that fires onLastClientGone() exactly once when the live connection count
drops to zero and stays zero for a configurable grace period. Arms only
after the first accepted connection (headless safety: CLI ao start never
self-stops). Reconnect before grace elapses cancels the pending timer.
Mutex guards liveCount/armed/pendingTimer; sync.Once guards the callback.
Tests use net.Pipe + a fake listener (no OS sockets); all three behavior
contracts verified with -race: never fires pre-connect, fires once on
last disconnect, reconnect within grace cancels and re-arms correctly.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(daemon): supervisor watchdog review fixes (data race, ErrClosed, leak, cleanup)
- Fix data race: capture liveCount into local inside lock before logging
- Replace io.EOF with errors.Is(err, net.ErrClosed) for correct production behavior; update fakeListener to match
- Fix goroutine leak: derive cancellable child context in Serve so watcher always unblocks on return
- Simplify makePipe: drop dead error return, update 3 call sites
- Remove dead defensive pendingTimer nil-check in armGrace
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(daemon): OS-native supervisor listener triggers clean shutdown
Creates platform-split Listen() in supervisor package (Unix UDS sibling
of run-file; Windows named pipe via go-winio). Wires it into daemon.Run
before srv.Run: listener failure is non-fatal so headless ao-start works.
Adds RequestShutdown() on Server and three unit tests for listen_unix.go.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(daemon): log supervisor Serve error instead of discarding it
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(desktop): supervisor link; daemon self-stops (clean) on frontend exit
Add connectSupervisor() in frontend/src/main/supervisor-link.ts: holds one
client connection to the daemon's OS-native supervisor socket (Unix domain
socket on macOS/Linux, named pipe on Windows). Retries with bounded backoff
until the daemon is up; reconnects automatically on drop.
In main.ts, connect the link from reportBoundPort (once per daemon ready
transition). Remove the quit-time daemon teardown: the before-quit handler
now only disposes the browser view. The process.on("exit") killDaemon call
is also removed. When Electron exits for any reason the OS closes the fd,
the daemon detects EOF, and self-stops after its 5s grace period. tmux and
ConPTY sessions survive and are adopted on the next boot.
killDaemon and stopDaemon are kept for the explicit user-stop path
(ipcMain.handle("daemon:stop", ...)).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(desktop): guard against daemon orphan when supervisor link is down; tidy test setup
Expose a live `connected` getter on SupervisorLinkHandle and add a
last-resort process.on("exit") kill that fires only when the link is
not actually connected (UDS never bound or addr was null), preventing
orphan daemons while keeping the OS-fd teardown path for the normal
case. Log a warning when addr is null so the skip is diagnosable.
Invert the setup.ts guard to the natural form, dropping the empty
if/else skeleton.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(core): restore promptless sessions in place (reboot recovery, no increment)
Drop the empty-prompt early return in restoreArgv that returned ErrNotResumable
when the adapter could not resume and no prompt was saved. Now control falls
through to GetLaunchCommand unconditionally: a saved prompt is replayed, an
empty prompt (orchestrator) launches fresh with the system prompt only, same id,
same workspace. Removes the only producer of ErrNotResumable and deletes the
now-dead sentinel, its service mapping, and the corresponding tests.
Frontend reference to SESSION_NOT_RESUMABLE in TerminalPane.tsx is left intact
(the handler becomes harmlessly dead; the API simply stops returning that code).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* test(integration): dead-live session is restored, not abandoned, after reconcile
Task 5 made promptless sessions relaunch fresh instead of hitting ErrNotResumable.
The reconcile crash-recovery path (documented in reconcileLive) terminates a
dead-live session then RestoreAll relaunches it on the same boot. This test
asserted the old promptless-stays-terminated artifact; update it to the intended
restored end state (live again, same id, one fresh runtime Create).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* style(session): replace em dashes in service.go messages and comment
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(desktop): scope supervisor link to spawn path; dispose link on explicit stop
Document that the liveness link is established only when the app spawned the
daemon: the attach path intentionally does not link, to keep an `ao start`
daemon persistent (headless safety). Also dispose the link on an explicit
daemon:stop so its reconnect loop does not retry a deliberately-stopped daemon.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
* docs: add daemon-lifecycle adopt-on-shutdown implementation plan
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
* fix(daemon): keep supervisor watchdog alive across transient Accept errors
Review of #2185: Serve previously returned on the first unexpected Accept error,
silently disabling the watchdog (the 'restart is caller's job' comment described
a contract the caller never fulfilled). Back off and keep accepting instead, so
a transient error (e.g. EMFILE) cannot leave the daemon unable to self-stop on
frontend death. Also drop the stale Task 3/4 planning comment.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(core): leave promptless workers terminated on restore (orchestrators still relaunch fresh)
A promptless, unresumable KindWorker session had no prompt to replay and
no native session id to resume from. Blank-relaunching it via GetLaunchCommand
would silently drop its task. restoreArgv now returns ErrNotResumable for this
case, gated on (meta.Prompt == "" && kind != domain.KindOrchestrator). Orchestrators
are promptless by design and continue to relaunch fresh with the system prompt only.
Re-introduces ErrNotResumable sentinel and its Conflict API mapping.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(desktop): re-link supervisor on attach for app-owned daemons (close lingering-daemon gap)
- runfile.Info gets Owner field (omitempty): "app" when Electron spawned,
empty for headless `ao start`. server.go reads AO_OWNER env to set it.
- daemonEnv() injects AO_OWNER=app so spawned daemons self-identify.
- Extracted establishSupervisorLink() from inline reportBoundPort code.
Spawn path calls it unconditionally; attach paths call it only when
shouldLinkOnAttach(owner) is true (owner === "app").
- Both attach paths (inspectExistingDaemon, resolveDaemonFromPort) now
read the owner from the run-file and re-link when app-owned, preventing
a lingering app-spawned daemon from self-stopping mid-session.
- Headless ao start daemons stay unlinked: persistent across app quit.
- New daemon-owner.ts + daemon-owner.test.ts (4 vitest cases, all pass).
- Go: 9 tests pass with -race; vet clean.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: quiet expected ErrNotResumable log in RestoreAll; note attach TOCTOU
Review polish on #2185: a promptless worker left terminated on boot is expected,
not an operational error, so log it at Warn not Error. Document the narrow
run-file re-read TOCTOU on the port-attach re-link path (worst case is linking a
headless daemon; the dispose-idempotency guard prevents any leak).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* style: gofmt the review-fix files (CI format/lint check)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* Zellij to Tmux and some other fixes. (#404)
* feat(runtime): add tmux adapter package
Adds backend/internal/adapters/runtime/tmux implementing ports.Runtime via
the tmux CLI. Drop-in replacement for the zellij adapter on Darwin/Linux.
Key design points:
- Handle is a plain session id string (no pane-id split needed for tmux).
- Exact-match session targeting via = prefix for kill-session and has-session.
- Keep-alive shell appended to launch command so sessions survive agent exit.
- send-keys -l chunked for literal text delivery (no key-name interpretation).
- IsAlive distinguishes definitive-dead (missing/no-server output) from probe
errors so the reaper never kills a session on a transient tmux failure.
- 34 tests pass: 32 unit tests via fakeRunner seam, 2 integration tests on
real tmux 3.6b (TestRuntimeIntegration, TestRuntimeIntegrationExactSessionParsing).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(tmux): address four code-review findings in tmux runtime adapter
- Remove em dash from tmux_test.go:462 (project hard rule); replace with semicolon
- Derive integration test session IDs from t.Name() so concurrent runs do not collide on the same tmux session
- Remove dead scaffolding variables (r/fr, r2/fr2) in TestCreateDestroysAndReturnsErrorWhenNotAlive
- Quote \${SHELL:-/bin/sh} in buildLaunchCommand and update all asserting tests
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(runtime): wire tmux on Darwin/Linux via runtimeselect, keep zellij on Windows
- New package runtimeselect: Runtime union interface (ports.Runtime +
SendMessage/GetOutput/AttachCommand) with compile-time assertions for
both adapters. New(log) returns tmux on non-Windows, zellij on Windows
(replicating the old daemon socket-dir setup).
- daemon.go: replace zellij-specific socket-dir block with
runtimeselect.New(log); update comment to be runtime-neutral.
- lifecycle_wiring.go: startSession param changed from *zellij.Runtime
to runtimeselect.Runtime.
- cli/doctor.go: runtime-aware checkTerminalRuntime (tmux on Darwin/Linux,
zellij on Windows); added checkTmux.
- cli/spawn.go: attach hint prints tmux attach -t <name> on non-Windows,
keeps zellij attach hint on Windows.
- wiring_test.go: startSession test uses runtimeselect.New(nil); zellij
direct tests retained for zellij-specific coverage.
- doctor_test.go: replaced three zellij tool tests with tmux equivalents.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: tidy runtime-neutral comments and doctor import grouping
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* refactor(tmux): drop unused runner.Start seam
tmux creates sessions detached via new-session -d, so the Start method
(carried over from the zellij runner shape, where it backs the Windows
fire-and-forget spawn) is never called. Remove it from the interface and
its implementations to shrink the seam.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(conpty): add protocol codec and output ring buffer (pure Go, OS-agnostic)
Ports the ConPTY named-pipe binary framing protocol and rolling output
buffer from pty-host.ts to Go. Implements EncodeMessage, MessageParser
(handles arbitrary chunk boundaries, payload copy guarantee), and Ring
(MaxOutputLines=1000, ANSI-safe, concurrent Append+Snapshot). All 15
unit tests pass on Darwin; GOOS=windows build is also clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(conpty): harden copy-safety and add concurrent ring test
Strengthen TestParserPayloadIsCopy to catch internal-buffer aliasing:
feed frame1, capture its payload, feed frame2 of the same length so the
parser's buffer overwrites the frame1 region, then assert frame1's bytes
are unchanged. The prior test only mutated the input slice post-Feed and
did not exercise the real aliasing risk.
Add TestRingConcurrent: 10 writer goroutines (Append) and 10 reader
goroutines (Snapshot + Tail) running concurrently with a WaitGroup. The
test is meaningful only under the race detector and catches any missing
mu coverage on Ring's exported methods.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(ptyregistry): port Windows pty-host sideband registry to Go
Adds package ptyregistry under backend/internal/adapters/runtime/conpty/ptyregistry.
Ports windows-pty-registry.ts: defensive read, atomic temp+rename write,
delete-on-empty, register-replaces-same-ID, and auto-pruning List.
PID liveness isolated behind build tags (syscall.Kill on Unix,
OpenProcess on Windows). 10 tests all green on Darwin.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(sdd): phase B briefs and progress for B1-B3
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(conpty): add pty-host serve engine with loopback TCP transport (B3)
Ports pty-host.ts behavior to Go: ptyConn interface seam, Serve engine
with ring replay, fan-out broadcast, MSG_* handlers, PTY-exit keep-alive,
and graceful shutdown (ConPTY dispose first, 50ms grace, then clients and
listener). Real conptyConn is Windows-only via build tag; non-Windows stub
keeps the package importable on Darwin/Linux. Tests use a fake ptyConn
with real loopback sockets and the B1 MessageParser, passing with -race.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(conpty): deliver scrollback snapshot and register client atomically
Review of Task B3 found one Important bug and two minors.
Important: in handleConn the ring Snapshot and the client registration
ran under two separate h.mu acquisitions. A PTY chunk arriving in that
gap was in neither the snapshot nor that client's broadcast, so it was
silently dropped (a hole in the client's stream). Now take the snapshot,
write it to the conn, and add the conn to the clients set all under a
single h.mu hold; broadcast also takes h.mu so it cannot interleave.
Added TestScrollbackLiveOrdering_NoDrop, which emits a contiguous
numbered stream while a client connects and asserts the client's stream
has no internal gap. It reliably fails against the old two-step code and
passes under -race -count=20.
Minor (faithfulness): conptyConn.Close() now also best-effort
Process.Kill() (nil-guarded) so a child that ignores ConPTY EOF still
exits and Done() fires, mirroring pty.kill() in pty-host.ts.
Minor (simplify): use os.Environ() instead of
exec.Command(shellCmd).Environ() for the child env.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(sdd): B4 brief
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(conpty): add runtime adapter with loopback pty-client and session management (B4)
Implements the conpty Runtime adapter: injectable spawn seam, loopback
TCP client helpers (SendMessage/GetOutput/IsAlive/Kill), and Runtime
methods (Create/Destroy/IsAlive/SendMessage/GetOutput). Session resolution
uses an in-memory map with B2 registry fallback for daemon-restart
recovery. Windows-only detached spawn in spawn_windows.go; stub errors
on other OSes. All adapter methods are unit-tested on Darwin against an
in-process B3 Serve and fakePTY. 48 tests pass, all three GOOS builds
succeed, vet clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(conpty): split IsAlive dead-vs-transient for reaper safety
clientIsAlive collapsed every probe failure (dial timeout, read-deadline
expiry, write error, connection-refused) to false, which the reaper turns
into ProbeDead and the LCM can promote to a permanent reap. A single
transient 2s loopback timeout would spuriously kill a live idle session.
Now clientIsAlive returns (alive bool, transientErr error): a refused dial
is definitively gone (false, nil); a timeout or any connected-then-failed
I/O error is transient (false, err) so the reaper records ProbeFailed and
retries. Wire IsAlive to propagate it. Add regression test covering both
the refused-is-gone and timeout-is-transient paths.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(sdd): B5 brief + ledger
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(terminal): stream-based Attach for tmux/zellij/conpty
Evolve the terminal layer from argv-based attach (PTYSource.AttachCommand
+ injected spawnFunc) to stream-based attach (Source embedding
ports.Attacher). tmux/zellij keep spawning their attach CLI on a local
PTY via the new shared ptyexec.Spawn; conpty attaches by dialing its
loopback pty-host directly with a loopbackStream over the B1 framing
protocol. Reattach/backoff/size/SIGWINCH/detach semantics are unchanged.
- ports: add Stream + Attacher.
- ptyexec: new shared package holding the creack/pty (unix) and ConPTY
(windows) spawn, moved verbatim from terminal with its tests.
- terminal: PTYSource -> Source, drop spawnFunc/WithSpawn, run loop calls
src.Attach and uses ports.Stream.
- tmux/zellij: add Attach (argv via ptyexec.Spawn); conpty: add Attach
(loopbackStream); ports.Attacher assertions on all three.
- runtimeselect: union embeds ports.Attacher in place of AttachCommand.
- tests migrated; new conpty attach_test against in-process Serve+fakePTY.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* style(ptyexec): replace em dashes carried from moved pty files
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(sdd): B6 brief + B5 ledger
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(runtime): select conpty on Windows, register pty-host subcommand, delete zellij
- runtimeselect.New: Windows branch now returns conpty.New(conpty.Options{}) instead
of zellij; compile-time assertion updated to conpty.Runtime.
- cli/ptyhost.go: new hidden "ao pty-host" subcommand (DisableFlagParsing so agent
shell args with leading dashes survive); calls conpty.RunHost and exits with its code.
- cli/root.go: wires newPtyHostCommand alongside newLaunchCommand.
- cli/doctor.go: Windows terminal-runtime check replaced with a static ConPTY
built-in pass; zellij import and checkZellij function removed.
- cli/spawn.go: Windows attach hint updated to dashboard message (ConPTY has no
CLI attach); zellij import removed.
- daemon/lifecycle_wiring.go: stale zellij comment updated to tmux/conpty.
- daemon/wiring_test.go: zellij import and TestDaemonZellijSocketDir test removed;
TestWiring_StartLifecycleThreadsMessengerIntoLCM now uses tmux.New.
- terminal/attachment_integration_test.go: re-pointed at real tmux
(TestAttachmentStreamsRealTmuxPane + TestAttachmentReattachAdoptsNewSize);
sessions cleaned up in t.Cleanup.
- internal/adapters/runtime/zellij: deleted entirely.
All three GOOS builds pass; go test -race ./... 1607 passed; go vet clean;
grep -rn "runtime/zellij" returns nothing.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(daemon): correct terminal-runtime comment to conpty on Windows
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(ptyexec): drop stale zellij reference in Windows spawn comment
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(sdd): final phase B ledger
* build(desktop): support local keychain signing for macOS builds
Bridge forge.config.ts to accept the local keychain flow (APPLE_SIGNING_IDENTITY
identity + AO_NOTARY_PROFILE notarytool profile) in addition to the existing CI
secrets path (CSC_LINK + APPLE_ID/app-specific-password). Enables a signed +
notarized macOS build from a developer Mac without exporting a .p12 or the Apple
ID app-specific password.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(daemon): default TERM so Finder-launched tmux attach works
A Finder/Dock launch starts the supervisor under launchd with no
controlling tty, so TERM is unset. The daemon inherits that, and its
tmux attach client (spawned with env=nil, inheriting the daemon env)
dies immediately with "open terminal failed: terminal does not support
clear" — the orchestrator terminal pane never opens.
Seed TERM=xterm-256color (what the renderer's xterm.js emulates) as the
base of buildDaemonEnv, the same place PATH is reconstructed for the same
class of "Finder launch lacks a terminal's env" bug. A real TERM from the
shell/process env still wins.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(lifecycle): plan for save-on-close/restore-on-open sessions
Captures the intended daemon lifecycle: on shutdown save every running
session (worker and orchestrator) plus its gitignore-respecting uncommitted
work to refs/ao/preserved/<id>, then force-remove worktrees; on boot recreate
worktrees, replay the preserved work, and restore all sessions. Reuses
existing SQLite state, session_worktrees.preserved_ref, manager.Restore, and
the /shutdown endpoint (no new file, migration, or route).
Also gitignore the built daemon binary copied into frontend/daemon/.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore(frontend): sync regenerated pnpm-lock and routeTree
Working-tree regeneration of the pnpm lockfile and TanStack Router generated
route tree. No hand edits; generated output only.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(workspace): add ForceDestroy for shutdown-path worktree removal
Adds ForceDestroy(ctx, info) to ports.Workspace and the gitworktree
adapter. It runs `git worktree remove --force`, then prune, then
os.RemoveAll as a backstop. A new worktreeForceRemoveArgs builder in
commands.go emits --force; the existing worktreeRemoveArgs is untouched
so Destroy still refuses dirty worktrees via ErrWorkspaceDirty.
TDD: test first creates a dirty worktree, confirms Destroy refuses with
ErrWorkspaceDirty, then confirms ForceDestroy succeeds and the path is
gone and deregistered. All 1609 backend tests pass.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(workspace): add StashUncommitted and ApplyPreserved for session lifecycle
Implements the correctness-critical save-on-close / restore-on-open pair
in the gitworktree adapter:
- StashUncommitted: captures uncommitted work (tracked edits and new
non-ignored files) via a temp GIT_INDEX_FILE into a real commit stored
at refs/ao/preserved/<session-id>. Never touches the real index or
stash stack. Returns empty string for clean worktrees. Logs the count
of .gitignore-skipped paths.
- ApplyPreserved: replays the preserve commit onto a freshly re-added
worktree via "git checkout <SHA> -- .". Deletes the ref on clean
success; keeps it and returns ErrPreservedConflict (wrapped) on
content conflicts.
- Adds both methods to ports.Workspace interface and stubs them in
integration and session_manager test doubles.
TDD: wrote two failing tests first (RED confirmed via build failure on
undefined methods), then implemented to GREEN. All 39 adapter tests pass.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(workspace): replace path-checkout with cherry-pick in ApplyPreserved
git checkout <sha> -- . is a path-checkout that always exits 0 for
content divergence, making ErrPreservedConflict unreachable. Replace
with git cherry-pick --no-commit which performs a true three-way merge,
leaves textual conflict markers on conflict, and exits non-zero so the
sentinel is correctly returned. Conflict detection now uses exit code
only (locale-independent). Add TestWorkspaceIntegrationApplyPreservedConflict
to assert: error is ErrPreservedConflict, preserve ref is kept, conflict
markers appear in the file. All 40 tests pass.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(session-manager): add SaveAndTeardownAll and RestoreAll for shutdown lifecycle
Implements Task 3: capture-then-destroy on shutdown and restore-all on startup.
- Adds ErrPreservedConflict to ports as a named sentinel; gitworktree aliases it
(following the same pattern as ErrBranchCheckedOutElsewhere).
- Extends the Store interface with UpsertSessionWorktree and ListSessionWorktrees
so the session manager can write the shutdown-saved marker and read it back.
- SaveAndTeardownAll: for every live session with a workspace path, stash
uncommitted work, write the session_worktrees row (DB commit before worktree
removal, crash-safety invariant), mark terminated, destroy runtime, force-remove
the worktree. Best-effort per session; no kind filter.
- RestoreAll: for every terminated session that has a session_worktrees row (the
marker written by SaveAndTeardownAll), re-create the worktree, apply any
preserved ref (conflict logs and continues), then relaunch via the existing
single-session Restore. Sessions killed by the user before shutdown (no row)
are skipped. Best-effort per session; no kind filter.
- TDD: 9 new tests (RED confirmed via build failure, GREEN confirmed 63 pass).
Full suite: 1621 tests across 77 packages.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(terminal): enable tmux mouse scroll and fix link clicking
On macOS the runtime is tmux, but two mouse interactions were broken in
the embedded terminal while copy/paste kept working:
- Scroll: the renderer drives scrolling by writing SGR mouse-wheel
reports into the pane (the zellij `--mouse-mode true` model), but tmux
ignores those reports unless mouse mode is on. Create only set `status
off`, never `mouse on`, so wheel scrolling silently no-opped. Enable
`set-option -t <id> mouse on`, mirroring the existing status-off step.
- Link clicking: the default WebLinksAddon handler calls window.open()
with an empty URL and then assigns location.href. Electron's
setWindowOpenHandler denies every window.open and only forwards the URL
passed to it, so the empty open is dropped and clicks no-op. Pass the
matched URL to window.open directly so the main process routes it to
shell.openExternal (the OS browser).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(session-manager): assert UpsertSessionWorktree precedes ForceDestroy
Add a shared ordered call log (sharedLog *[]string) to both fakeStore
and fakeWorkspace. TestSaveAndTeardownAll_CaptureOrderAndMarker now
wires both fakes to the same slice and asserts upsertIdx < forceIdx,
enforcing the crash-safety invariant that the DB write is committed
before the worktree is force-destroyed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(daemon): wire RestoreAll/SaveAndTeardownAll into boot/shutdown sequence
Exposes session manager through a minimal sessionLifecycle interface
(RestoreAll, SaveAndTeardownAll) returned from startSession, then calls
RestoreAll (best-effort) before srv.Run and SaveAndTeardownAll with a
fresh 30s-bounded context after srv.Run returns. Both SIGTERM and POST
/shutdown funnel through srv.Run returning, so the single save call site
covers both paths.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(daemon): fix seam-test tautology and lifecycle variable shadow
Finding 1: dispatch both sessionLifecycle methods through an interface
variable (var sl sessionLifecycle = fake) so the runtime body exercises
interface dispatch, not just direct struct method calls.
Finding 2: rename local variable 'lifecycle' to 'lc' in
TestWiring_StartSessionBuildsSessionService to remove the shadow of the
imported lifecycle package.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(frontend): call POST /shutdown before killing daemon on quit
In before-quit, POST /shutdown (8s AbortSignal.timeout) so the daemon
saves sessions gracefully before the SIGTERM kill. Adds a re-entrancy
guard (quitting flag) so a concurrent app.quit() cannot double-preventDefault.
Falls back to killDaemon on fetch failure or timeout: quit is never blocked.
Keeps the process.on('exit') SIGTERM fallback intact.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(storage): guard session_worktrees.state against empty-string CHECK violation; add ponytail comments
The save path (saveAndTeardownOne) never sets domain.SessionWorktreeRecord.State,
so it arrives at UpsertSessionWorktree as "". The generated upsert includes state
in the INSERT column list, so the DB default ('active') is never applied and the
CHECK constraint (state IN ('active', ...)) would fire at the first real shutdown.
Fix: default to 'active' in the store adapter when row.State is "". No schema
change, no migration, no gen edit.
Also add ponytail: comments on the State field (domain type), the write path, and
the read path, documenting that state is unused multi-repo scaffolding and that the
upgrade path is to wire a real value when multi-repo worktree lifecycle states ship.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(storage): add real-SQLite test for empty-State guard in UpsertSessionWorktree
Adds TestUpsertSessionWorktreeEmptyStateDefaultsToActive to the store
test file. It inserts a SessionWorktreeRecord with State at zero value
"" via UpsertSessionWorktree against a real SQLite DB, then reads the
row back and asserts State == "active". This directly exercises the
guard added in the prior commit and would fail if the guard were
removed (the CHECK constraint rejects ""). Mirrors the helpers and
setup pattern of TestSessionWorktreesRoundTrip exactly.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(comments): correct shutdown-mechanism and task-ref inaccuracies
Fix 1: daemon.go comment near SaveAndTeardownAll now correctly states
that POST /shutdown closes the shutdownRequested channel (not cancel ctx).
Also tighten the RestoreAll comment to remove the inaccurate claim.
Fix 2: remove "Task 2's" phrasing from ForceDestroy ponytail comment in
workspace.go; condition still references StashUncommitted by name.
Fix 3: add note in main.ts that the 8s fetch timeout is shorter than the
daemon's 30s save bound, so a SIGTERM after fetch abort does not cut the
in-flight save short.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: remove .superpowers workflow scratch from repo
These SDD workflow artifacts (task briefs, agent reports, progress ledger,
review packages) were committed by accident in prior work, against the
.superpowers/sdd/.gitignore intent. Remove them from the repo; they remain
local-only scratch.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(spec): graceful restore + post-failure orchestrator recreate
Fix the opaque 500 when restoring an un-resumable session (typed 409
SESSION_NOT_RESUMABLE), and add a post-failure popup that offers to recreate a
fresh orchestrator on the same branch (cleaning the worktree, preserving
committed history). Orchestrators only; recreate fires only after a restore
attempt confirms the session cannot be resumed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(plan): restore-recreate orchestrator; reuse existing /orchestrators clean=true
Planning discovery: the recreate capability already ships via POST /orchestrators
(clean=true), which kills the dead orchestrator and re-spawns on the canonical
branch (addWorktree reattaches an existing branch). So the feature collapses to a
typed-error fix plus a frontend popup. Spec updated to match.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(session): return typed SESSION_NOT_RESUMABLE instead of 500 on un-resumable restore
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(renderer): offer recreate-orchestrator popup when a session cannot be restored
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(spec): drop stale OpenAPI-regen note (feature adds no route)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(ci): gofmt/goimports, golangci-lint hygiene, and Windows-aware doctor tests
Formatting: ran gofmt and goimports (with local-prefixes) on the 8 listed
files plus ptyexec/spawn_unix.go which the linter also flagged.
Lint (25 issues fixed):
- gosec G115: EncodeMessage now returns ([]byte, error) with an explicit
bounds check before the int->uint32 conversion; all callers updated.
- govet nilness: removed dead `if lastErr == nil` branch in clientIsAlive;
lastErr is provably non-nil at that point (real bug).
- nilerr: extracted runAcceptLoop helper so Accept-error-on-close is not
flagged; listener close is normal shutdown, not a caller error.
- staticcheck SA4010: removed dead `full = append(...)` loop in host_test.
- revive var-declaration: `var prev int = -1` -> `prev := -1`.
- revive redefines-builtin-id: deleted local `min` helper; builtin covers it.
- unparam (2): dropped always-nil env return from attachCommand; dropped
unused shellPath param from buildLaunchCommand; updated callers.
- errcheck (8): deferred Close/Remove calls wrapped in func(){_ = ...}();
type assertion in host_main.go uses ok-form; fmt.Fprintf to stdout uses
_, _ = pattern; workspace.go tmpIdx.Close() uses _ =.
- gocritic nestingReduce: inverted if+continue in runtime.go resolve loop.
Windows E2E: skip TestDoctorChecksTmuxVersion,
TestDoctorChecksTmuxVersionFailsOnError, TestDoctorWarnsWhenTmuxMissing on
windows (ao doctor emits a conpty check there, not tmux).
Verified: gofmt -l . clean, golangci-lint 0 issues, go build ok,
go test -race 1624/1624 pass.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(ci): set git identity in worktree clone fixture; loosen tmux reattach timeouts
The preserve round-trip/conflict tests commit inside a worktree of the cloned
repo, which had no git identity; CI runners cannot auto-derive one, failing with
"empty ident name". Set user.email/user.name on the clone in setupOriginClone so
its worktrees inherit it.
The tmux reattach test drives a real shell and parses stty output, which is slow
under -race on CI; raise its echo-write and SIZE-output waits.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(terminal): resend size probe on tmux reattach until the shell answers
Bumping timeouts was the wrong fix: a 30s wait still failed, so the probe output
deterministically never appeared, not slowness. onOpen signals the stream accepts
input, not that the reattached sh -i is at a prompt, so the first echo keystroke
can be dropped. Resend the probe each poll until SIZE output lands, and on timeout
dump the captured pane buffer so a remaining failure is self-explaining.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(terminal): set TERM for real-tmux attach tests so they run in CI
Root cause (from the buffer dump the prior commit added): with TERM unset on CI
runners, tmux refuses to attach a client and prints "open terminal failed:
terminal does not support clear", so the pane never runs the size probe. The
daemon defaults TERM in production; the tests bypass it. Set TERM=xterm-256color
in both real-tmux tests. Reproduced locally with `env -u TERM` (fails the same
way) and verified the fix passes under it.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(spec): crash-proof session reconcile design
Boot-time reconcile makes live tmux + worktree state match the DB on every
daemon start, so a SIGKILL/crash/force-quit that skips SaveAndTeardownAll no
longer leaks an orphaned daemon, tmux sessions, or worktrees. Adopt
crash-surviving tmux sessions, preserve-and-terminate dead ones, reap
in-namespace orphans, and add a frontend kill+replace branch for a wedged
orphan daemon.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(spec): simplify reconcile to per-session IsAlive, drop ListSessions
Every leak in the incident maps to a DB row, so orphan-reap is a per-session
IsAlive+Destroy over terminated rows; no runtime enumeration, no ports/conpty/
runtimeselect changes. Reaping a tmux session with no DB row is deferred (YAGNI).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* docs(plan): crash-proof session reconcile implementation plan
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* feat(session): reconcile live pass (adopt alive, stash+terminate dead)
* feat(session): reconcile reap pass and Reconcile entry point
* feat(daemon): run Reconcile on boot in place of bare RestoreAll
* test(integration): reconcile terminates dead-live sessions and reaps leaked tmux
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* test(integration): correct misleading CreateSession comment in reconcile test
* feat(frontend): kill+replace a wedged orphan daemon on launch
When both inspectExistingDaemon and resolveDaemonFromPort return null but
a process still holds the daemon port (a crashed/orphaned daemon), spawning
a new Go child would collide on the port and exit 1. Detect this case, SIGTERM
the holder (via the run-file PID, falling back to the probe PID), poll until the
port is free (up to 8s), clear the stale run-file, then proceed to spawn fresh.
The healthy-daemon reuse path is unchanged.
Pure helper: src/shared/daemon-takeover.ts (planDaemonTakeover)
Unit tests: src/shared/daemon-takeover.test.ts (3 tests, TDD red-green)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(frontend): fire orphan-daemon takeover when a holder actually exists
Replace planDaemonTakeover (inverted logic: ran kill block only when probe
was null) with shouldReplacePortHolder(probe, holderPidAlive) which returns
true when a real holder exists: non-null probe (rejected responder) OR a
run-file PID that is still alive (hung holder). Update main.ts call site to
compute PID liveness before gating the kill block. Update tests to cover all
three distinct outcomes non-vacuously.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* docs+test: accurate takeover comments, reconcileLive probe-error test, Reconcile doc
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(session): restore promptless orchestrators and crash-orphaned sessions
The orchestrator was abandoned on every app open: a fresh orchestrator
spawned each launch and the prior conversation appeared lost (it was not;
the transcript stays in ~/.claude, resumable by the deterministic
--session-id AO pins). Two defects combined:
1. Restore's guard rejected any session with no agentSessionId AND no
prompt as ErrNotResumable. But Claude resumes via a deterministic
session id regardless of those fields, so promptless orchestrators
were perfectly resumable yet always rejected. Workers slipped through
only because they carry a prompt. Move the resumability decision to the
adapter: restoreArgv returns ErrNotResumable only when GetRestoreCommand
reports it cannot resume AND there is no prompt to fresh-launch from.
2. reconcileLive marked a crash-orphaned (dead-runtime) session terminated
without a restore marker, so RestoreAll skipped it and it stayed dead.
It now saves-and-tears-down to the same end state a graceful shutdown
produces (capture work, write the session_worktrees marker, terminate,
remove the worktree), so RestoreAll relaunches it on the same boot,
resuming history. Crash recovery now matches graceful restart. If work
capture fails it terminates without a marker rather than risk losing
un-preserved work.
Tests: promptless orchestrator restores via adapter resume; promptless
session with a non-resuming adapter still returns ErrNotResumable;
reconcileLive writes the marker + tears down the worktree. Full backend
suite green (1632), gofmt/vet clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
(cherry picked from commit 9ae05735d6f06ac989857534bae2766392772c71)
* chore: format with prettier [skip ci]
* docs: scrub stale zellij references after tmux/conpty migration (#409)
PR #404 migrated the runtime adapter from Zellij to tmux (Darwin/Linux)
plus conpty (Windows), selected via runtimeselect, but ~30 stale zellij
references lingered in comments and docs describing zellij as the current
runtime. This is a comments/docs-only cleanup with no behavioral change:
comments now say tmux (or tmux/conpty when both platforms are relevant),
terminal/doc.go and docs/backend-code-structure.md are rewritten to
reflect the tmux + conpty + runtimeselect attach model, and the daemon
environment, STATUS, stack, architecture, and CLI docs are updated.
Also gitignore the local .codegraph/ and .cursor/ tooling dirs.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Joins the ground-up ReverbCode rewrite (this tree) with the prior
agent-orchestrator history as a second parent, preserving both lineages.
Tree content is ReverbCode's; old history remains reachable in main.
* fix(desktop): package Windows via NSIS instead of Squirrel (#401)
Squirrel.Windows is a poor fit: per-user install only, no custom install
directory, no proper add/remove-programs uninstaller, and fragile updates.
Replace it with a real NSIS installer (per-user or per-machine, custom
install dir, uninstaller), matching recordly's working Windows setup.
Electron Forge ships no first-party NSIS maker, so add a thin MakerBase
subclass (makers/maker-nsis.ts) that bridges to electron-builder's
buildForge, the same engine electron-builder uses, scoped to win32. The
maker exposes the NSIS knobs the issue calls for (oneClick:false,
allowToChangeInstallationDirectory, per-machine) and defaults to an
assisted installer.
- forge.config.ts: drop maker-squirrel, add the NSIS maker instance.
- testing-build.yml: target "nsis"; smoke-install via /S under out/make;
drop the Squirrel-specific log capture.
- Rename the package "agent-orchestrator-frontend" -> "agent-orchestrator":
this repo is the full app, not just a frontend. User-facing naming was
already "Agent Orchestrator" (productName) / agent-orchestrator.exe.
Deferred (per issue, separate follow-ups): bundling zellij.exe so a fresh
Windows install needs no manual zellij, an actionable "zellij not found"
error at session-create, and Windows code-signing.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
* fix(desktop): disable electron-builder publish + drop electron-squirrel-startup
CI fix: the NSIS maker's electron-builder run inferred a GitHub publish
target from package.json `repository` and tried to upload (to emit
auto-update info), failing with "GitHub Personal Access Token is not set".
Forge owns publishing (the workflow uploads via `gh release`), so set
`config.publish = null` to disable electron-builder's upload entirely.
Also remove `electron-squirrel-startup`: it only handled Squirrel.Windows
install/update hooks (--squirrel-* flags) and is dead weight under NSIS.
Drop the dependency, its import, the startup quit-block, the whenReady
guard, and the type shim. The EPIPE std-stream guard stays (it covers any
windowless Windows GUI launch, e.g. from a shortcut).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: gitignore electron-builder's builder-debug.yml dump
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* fix(desktop): handle Squirrel startup events + CI smoke-install the Windows build
Two Windows-install fixes.
1. Add electron-squirrel-startup handling. main.ts had no handler for the
--squirrel-{install,updated,uninstall,obsolete} flags Squirrel runs the exe
with during install/update/uninstall. Without it the install hook booted the
full app (window + daemon spawn) and never exited, so the installer hung
waiting on it. Now we create/remove shortcuts and quit immediately; it is a
no-op on macOS/Linux.
2. Add a Windows CI smoke-install step. After the build, run Setup.exe --silent
on the clean native x64 windows-latest runner, assert the install dir is
created, and upload SquirrelSetup.log as an artifact. This captures the
install log we otherwise cannot get and gives a build-vs-host verdict: a clean
install proves the artifact is good, so a failing user machine is host-side
(AV/disk/signing). The runner has no real-time AV, so it does NOT prove
SmartScreen/Defender will accept the unsigned binaries on end-user machines;
code-signing stays the durable fix.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* ci(windows): capture SquirrelSetup.log from the app dir, fix misleading warning
Update.exe writes SquirrelSetup.log into %LocalAppData%\AgentOrchestrator, not
SquirrelTemp; the smoke step looked only in SquirrelTemp and so printed
"failed before Update.exe ran" even on a successful install (exit 0, dir
created). Look in the app root dir first, fall back to SquirrelTemp, and drop
the false-failure wording.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
On Windows the daemon shells out to console-subsystem processes without
suppressing the console window. The reaper polls session liveness on a
timer (and the attach loop retries), each call runs a zellij command, and
every invocation pops a console window that instantly closes, so the user
sees rapid blinking.
- Add a platform-split hideWindow(cmd) helper to the zellij package
(CREATE_NO_WINDOW + HideWindow on Windows, no-op elsewhere) and call it
in execRunner.Run so list-sessions and friends stop flashing.
- startBackgroundProcess now uses CREATE_NO_WINDOW instead of
CREATE_NEW_CONSOLE (the latter creates a console then hides it, flashing).
- Pass windowsHide: true to the daemon spawn in main.ts so the daemon's own
console stays hidden on a Windows GUI launch.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* fix(desktop): recover login-shell env so the daemon finds zellij/credentials
A Finder/Dock launch starts the app via launchd, not a login shell, so
~/.zprofile and ~/.zshrc are never sourced. The daemon then inherits
launchd's minimal env (no /opt/homebrew/bin on PATH, no exported
ANTHROPIC_API_KEY, etc.), cannot exec zellij/git, and its agents cannot
authenticate. Launching from a terminal masked this because the shell had
already populated the env, so it only reproduced on a real Finder/Dock launch.
Resolve the login-shell environment once at startup
($SHELL -ilc "printf sentinel; env -0"), adopt it as the base for
daemonEnv(), and force PATH from the shell with a static floor when the probe
fails (timeout/non-zero exit). The probe never blocks startup (3s SIGKILL
timeout, stdin closed) and degrades to the floor rather than erroring.
Windows keeps its existing behavior: its env comes from the registry/session
and is inherited by GUI-launched apps, so this bug does not exist there.
Pure parse/merge logic lives in shared/shell-env.ts (no node:* imports, per
the daemon-attach.ts convention) with unit tests; main.ts owns the real spawn.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: format with prettier [skip ci]
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* feat(cli): add examples to ao preview --help (#387)
Update the preview command help to include a concise Examples section
covering the four common workflows:
ao preview
ao preview file:///home/aoagent/ReverbCode/index.html
ao preview http://localhost:5173
ao preview clear
Replace the workspace-relative path (./dist/index.html) in the Long
description with the absolute file:// URL pattern agents actually use.
Command syntax and behavior are unchanged.
Closes#387
* fix(cli): clarify no-arg preview fallback behavior in help text
Address review feedback on #388: the no-argument description now
reflects the daemon's actual behavior — autodetect the workspace
entry point, fall back to the session's existing preview target
when none exists.
---------
Co-authored-by: AO Bot <ao-bot@composio.dev>
On Windows Squirrel/GUI launches there is no attached console, so
process.stdout/stderr are dead pipes. The daemon-output console.log/error
calls failed with EPIPE and, lacking an error listener, crashed the main
process with 'A JavaScript error occurred in the main process'. Add an
error listener that ignores broken-pipe writes.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Replaces linux/windows/macos-testing-build.yml with a single testing-build.yml
(matrix over ubuntu/windows/macos) that publishes all artifacts to one
0.0.0-testing-<sha> prerelease. Manual dispatch + 0.0.0-testing-* tag push.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Harshit Singh Bhandari
neversettle
Apoorv Singh
NIKHIL ACHALE
Dushyant Singh Hada
swyam sharma
Adil Shaikh
Vaibhaav Tiwari
PRADEEP SAHU
Priyanshu Choudhary
prateek
Laxman
i-trytoohard