[namastex888]

Summary

The user confirmed there is no Genie UI any more — the only "UI" is the Orca integration. So genie ui-bridge, the stdio bridge built for the separate-repo dash fork, is retired on exactly the terms genie mcp was in #2820: the command becomes a stub that writes a stable diagnostic to stderr and exits 1, and everything that existed only behind it is deleted.

$ genie ui-bridge
Error: genie ui-bridge has been retired; the Orca integration is the supported UI surface, or roll back to a pre-retirement signed release.
$ echo $?
1

22 files changed, 86 insertions(+), 4388 deletions(-).

Removed

File	LOC
src/lib/v5/mcp-server.ts + .test.ts	390 + 269
src/lib/v5/mcp-tools.ts + .test.ts	1126 + 1313
src/lib/v5/bridge-watcher.ts + .test.ts	139 + 173
src/lib/v5/UI-BRIDGE.md	152
src/term-commands/ui-bridge.ts 295 → 14, .test.ts 486 → 47	−720

Nothing outside that set imported them. src/lib/v5/authority-barriers.test.ts was the one external consumer (openReadonlyDb from mcp-tools.ts); its two assertions are dropped because the surviving readonly open, openReadonlyHandle, is asserted on the same line of the same tests — the barrier keeps identical coverage.

Two exports went dead with the tool registry and were removed too: listWishSlugs and WishGroupRow in task-state.ts. The SessionStart hook bundle keeps its own listKnownSlugs, so branch resolution is unaffected. knip.json also drops ss from ignoreBinaries — it was there only for the old "the bridge never opens a socket" test.

Kept, and why

• openWithWalIndexRecovery and its lock machinery in sqlite-open.ts. Its only caller was tryWriteOpen in mcp-tools.ts, and with that gone the component that creates the poisoned WAL index is gone too — so no shipped path calls it and no shipped path needs it. It stays because it is a ~250-line hardened cross-process-locked recovery seam with its own direct test suite, and it is the documented opt-in for any future single-owner writer that reintroduces a degraded read-only session. Its comments now say exactly that instead of pointing at deleted files. Removing it is a clean follow-up if reviewers prefer.
• hireAgent / unhireAgent / getHire / listHires in task-state.ts. The bridge was their only shipped writer, but hire_roster rows are still part of genie task export/import state (roadmap sync excludes them as machine-local) and v5-task.test.ts seeds through them. Section comment updated to record that the writer is gone.

Reference cleanup

• src/lib/interactivity.ts — ui-bridge stays in the workspace-gate bypass set for the same reason mcp does: the legacy gate must not exit 2 and mask the retirement diagnostic. Comment rewritten.
• CLAUDE.md — CLI table row now reads "Retired"; the "product MCP server … is retired" gotcha now says the UI bridge is retired too rather than "is a separate private stdio bridge and is NOT the retired product integration".
• README.md — command table row and the MCP-retirement section.
• src/__tests__/claude-md-drift.test.ts — ui-bridge dropped from REQUIRED_V5_COMMANDS, matching mcp's absence: a retirement stub is not the shipping surface.
• Stale prose in identity.ts, resolve-wish-branch.ts, genie-db.test.ts, sqlite-open.test.ts that pointed at the deleted modules.
• .genie/INDEX.md and .genie/roadmap.json mention genie-ui-bridge only as SHIPPED historical ledger entries, so they are untouched. grep over skills/, plugins/, scripts/ and docs/ found no ui-bridge references at all — nothing to report in the docs submodule.

Conflict note

The open PR fix/orca-mode-guards edits src/term-commands/ui-bridge.ts's Orca-mode error path. That work is superseded by this deletion — the stub opens no database, so there is no Orca-mode barrier left to guard on this path. Whichever lands second will need a trivial conflict resolution in favour of the stub.

Evidence

Run in a clean worktree off origin/dev (a13384ea0), umask 022:

• bun run typecheck — clean
• bun run lint — exit 0, 3 pre-existing complexity warnings (doctorCommand 37, parseJsonRejectingDuplicateKeys 29, codexPluginSurfaceChecks 27), unchanged from baseline
• bun run dead-code — clean (baseline was clean; the two new knip findings this change surfaced were fixed by deleting the dead exports, not by ignoring them)
• bun run lint:complexity-budget — OK: budget intact
• bun run lint:hook-content — hook-content-binding: OK
• bun run build — bundles 305 modules

Targeted tests (the full suite OOMs on this box):

File	Result
src/term-commands/ui-bridge.test.ts, mcp.test.ts, src/__tests__/claude-md-drift.test.ts, src/lib/v5/resolve-wish-branch.test.ts	40 pass / 0 fail
src/lib/v5/authority-barriers.test.ts, src/lib/v5/task-state.test.ts	125 pass / 0 fail
src/lib/v5/sqlite-open.test.ts, src/lib/v5/genie-db.test.ts	39 pass / 0 fail
src/term-commands/v5-task.test.ts	76 pass / 0 fail
scripts/release-docs.test.ts	44 pass / 0 fail

The new ui-bridge.test.ts mirrors mcp.test.ts and adds one assertion the MCP stub does not have: run in an empty tmpdir with a fresh GENIE_HOME, the command must leave the directory completely empty — no .genie/, no database.

🤖 Generated with Claude Code

https://claude.ai/code/session_013gGxGgKskzyzr1HRUB6cV3

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: 24b0f63a-a5b0-4a88-b6cb-2fbfce17698b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

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.}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-08-30T01:29:46.359755Z	b2661cd	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b2661cd184

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Remove the newly dead WAL recovery machinery

Because this commit deletes mcp-tools.ts and its tryWriteOpen caller, these lines confirm that no shipped path calls this recovery code or creates the condition it handles; only its tests now exercise the cross-process lock and recovery implementation. Retaining substantial stateful machinery solely as a seam for a hypothetical future writer violates the repository's KISS release gate, so delete the recovery implementation/export and its tests along with the last caller, or identify a present contractual consumer.

AGENTS.md reference: AGENTS.md:L27-L30

Useful? React with 👍 / 👎.