[milind-soni]

Release 0.1.122's CI run (39684923366, on 087f01f31) failed on every OS. Every failure was a test that hadn't caught up with the product, a Linux-only suite running on Windows, or an overloaded shard. This PR changes no product code.

Every OS

File	Decision	Why
electron/updater.test.mjs	test	ensureCloudAccount's onState also tells includedPages (#3266). The vm world the test builds had no includedPages, so it threw a ReferenceError. It now gets the same stub as includedX.
scripts/docker-image.test.ts	test	The Cloud home's tools layer now installs a pinned pnpm@10.33.0 (bot box commit), and cleans up after it in the same step. The test now counts 3 heavy installs for cloud-home (still beneath the app files), reads every rm -rf in a step (apt's lists come first), and checks ENGINES_CACHE_TAG on the engines' install, not on pnpm's. Mutation-checked: removing pnpm's cleanup fails the cache test.
server/cloud-send-move.e2e.test.ts	test	A send of some items (keys) never carries chats (withChats = !keys && …), and S4 deleted Bring. So nothing put Kai's and Lux's lines on My Cloud before the sender strip, and My Cloud held no rows at all. The test now uses an older app's whole Send everything, which did carry them, and asserts that it did. The whole file passes locally (5/5).

Windows only

File	Decision	Why
server/bot-box/{plan,bridge,broker,egress-gate,harness-gate}.test.ts	platform gate	The bot box runs only on a Linux Cloud home (detect.ts returns not-linux anywhere else). Its gates listen on unix sockets, which Windows refuses at a file path (EACCES). Its plan is bubblewrap argv over POSIX paths. The suite is describe.skipIf(win32), with the reason in a comment.
server/bot-box/box.test.ts, broker-mcp.test.ts	platform gate, per test	Only "on" / lend / the two broker tests skip. "off" (what Windows actually gets) and the plain wrap tests still run there. afterEach no longer removes a folder that was never made.
server/workspace-backup.test.ts	platform gate, socket only	On Windows the box/ folder holds only the seccomp filter, with no live socket, and the test still checks the folder is left out of the backup.
server/cloud-team.e2e.test.ts (7th, seen once the relaunch worked)	test	The prompt names the memory file the way workspace.ts quotes it (JSON.stringify), so on Windows its backslashes come doubled. The test now looks for the quoted path.
server/cloud-team.e2e.test.ts (6)	fixture	On Windows, kill("SIGTERM") ends the team server without running its shutdown. sessions.json.open stays behind, so the next start drops every email session as it would after a crash (by design). All 6 failures followed from that relaunch. scripts/testing/cloud-send-fixture.ts now stops the team server over IPC on Windows, and the server runs its own SIGTERM shutdown (the same approach as acp-shutdown.e2e.test.ts). Checked on macOS both ways: forcing the IPC path passes 7/7, and hard-killing gives exactly the 6 CI failures.
shared/app-frame/host-page.test.ts	.gitattributes	app-host.html is now pinned eol=lf. A CRLF checkout broke both the byte-for-byte compare with the build script's output and the hash on its first line.
electron/shared-drives.node-test.mjs (from #3254/#3272, after the release)	test	On Windows, "/" is the current drive's root. On CI that is D:\ (the checkout), a separate data drive from the home folder's C:\, which may be shared whole by design. The "system disk" assertion now uses the root of the home folder's drive.
server/routes/published-pages.test.ts	test	The guest path maps to this computer's native path (join), so the expected value is join("/vm/bot-1", "site").
scripts/gen-app-fixtures.test.ts	test	Written paths are compared relative to the root with /.
server/app-docs.test.ts	test	The 0o600 check skips Windows, as the repo's other mode checks do.

Shards

scripts/testing/vitest-shard-weights.json was last refreshed on Oct 12. It had no entry for one-home.e2e (215 s), routine-recovery.e2e (168 s), cloud-computer-auto-sleep.e2e (106 s) or 140 other e2e files, and every one of them landed in shard 4. That shard's Linux job hit its 30-minute cap and was cancelled on main (39684885971, 39687280672), as well as on this branch's first dispatch. I regenerated the file with update-shard-weights.mjs --run 39684923366 --run 39684885971. Using this PR's Linux per-file times, the test seconds per shard are now 1133/1085/1090/1159, down from 942/958/1000/1566. No timeout was raised.

No ratchets touched (server/index.ts, electron/main.mjs unchanged), no locale files.

Local (macOS): the touched files plus team-move.e2e give 20 files / 143 tests passing (the Linux self-check skips), plus typecheck, lint and the ci-scope tests.

🤖 Generated with Claude Code

Summary by CodeRabbit

• Tests
  • Improved cross-platform test reliability, especially on Windows, with platform-aware handling for paths, permissions, process shutdown, and temporary resources.
  • Skipped Linux- and Unix-specific integration tests on Windows where required.
  • Updated Docker image and Cloud fixture coverage to reflect installation and cleanup behavior.
  • Expanded send-operation coverage to verify multiple conversations are included.
  • Adjusted test distribution weights and fixture path checks for more consistent test runs.
• Chores
  • Standardized line endings for a shared application page.

[claude]

⚠️ Code review skipped — your organization has no extra usage available to pay for this review.

If your organization's extra usage balance is empty, an organization admin can add extra usage credits at claude.ai/admin-settings/usage. If its monthly spend limit was reached, an admin can raise it on the same page. If neither applies, contact Anthropic support.

Once extra usage is available, someone with write access to this repository can comment @claude review on this pull request to trigger a review.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
openmausbot-docs	Ready	Preview	Oct 22, 2026 1:12am UTC

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: defaults
• Review profile: CHILL
• Plan: Advanced
• Run ID: f4263d06-fbf6-4d27-9021-633cdda705ff

📥 Commits

Reviewing files that changed from the base of the PR and between f12d03f and 2700bc9.

📒 Files selected for processing: 2

• scripts/testing/vitest-shard-weights.json
• server/cloud-team.e2e.test.ts

Included review availability: This review used your included allowance. Your plan provides up to 8 included reviews per hour; 3 remain after this review.

📝 Walkthrough

The changes update Windows test handling, platform-aware path expectations, Team Cloud fixture shutdown, Docker image assertions, and Vitest shard weights. They also adjust cloud-send and prompt-path test expectations and add an LF checkout rule for app-host.html.

Changes: Cross-Platform Test and Harness Updates

Layer / File(s)	Summary
Platform-specific test execution server/app-docs.test.ts, server/bot-box/*test.ts, server/workspace-backup.test.ts	Tests that require POSIX features skip on Windows. The workspace backup test creates and closes its broker socket only on non-Windows platforms.
Platform-aware paths and expectations .gitattributes, electron/updater.test.mjs, electron/shared-drives.node-test.mjs, scripts/gen-app-fixtures.test.ts, server/cloud-send-move.e2e.test.ts, server/cloud-team.e2e.test.ts, server/routes/published-pages.test.ts	Tests use platform-aware paths and roots. The updater mock includes an onState handler. The cloud-send test checks transferred chat messages, and the prompt-path test expects JSON-escaped workspace paths. .gitattributes sets LF checkout for app-host.html.
Team Cloud fixture shutdown scripts/testing/cloud-send-fixture.ts	On Windows, the fixture starts the team process with IPC and sends a stop message during relaunch. It waits for exit when connected and uses SIGTERM when disconnected.
Docker and test-shard assertions scripts/docker-image.test.ts, scripts/testing/vitest-shard-weights.json	Docker tests account for the additional Cloud home install, inspect cleanup operands across install commands, and check engine cache tagging on the engines install. The shard-weight map updates weights and test entries.

Priority: ⬇️ Low

Estimated code review effort: 3 (Moderate) | ~20 minutes

Change: Other

Sequence Diagram(s)

sequenceDiagram
  participant CloudSendFixture
  participant TeamCloudProcess
  participant IPCConnection
  CloudSendFixture->>TeamCloudProcess: Start process with IPC enabled
  TeamCloudProcess->>IPCConnection: Register stop handler
  CloudSendFixture->>IPCConnection: Send stop message on relaunch
  CloudSendFixture->>TeamCloudProcess: Wait up to 10 seconds for exit

Loading

Suggested reviewers: claude

Merge Risk: ⚪ Minimal · up to 2700b

No concrete issue remains that would block merging these test and shard-weight updates after normal checks.

Pre-merge checks: 4 · 1

Failed checks: 1 warning

Check name	Status	Explanation	Resolution
Docstring Coverage		Docstring coverage is 20.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 5 functions across 17 files. (1 skipped: …	Write docstrings for the functions missing them to satisfy the coverage threshold.

Passed checks: 4 passed

Check name	Status	Explanation
Title check		The title clearly summarizes the primary changes: restoring green CI by updating tests and excluding Linux-only box tests on Windows.
Description check		The description provides detailed change reasons, platform-specific context, shard updates, and verification results. It does not reproduce the template headings or checklist, but the required informa…
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

Full details: Description check

Explanation

The description provides detailed change reasons, platform-specific context, shard updates, and verification results. It does not reproduce the template headings or checklist, but the required information is mostly present and the changes are clearly documented.

Full details: Docstring Coverage

Explanation

Docstring coverage is 20.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 5 functions across 17 files. (1 skipped: 1 unsupported.)

Resolution

Write docstrings for the functions missing them to satisfy the coverage threshold.

✨ Finishing Touches: 💡 1

📝 Generate docstrings: 💡

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}