Run all available tests and linters, check for anti-patterns, and generate a comprehensive code quality report
The English summary is being prepared. Showing the original description.
Check requirements docs against implementation and report discrepancies
The English summary is being prepared. Showing the original description.
See how to installReview the instructions the agent will receive before you install.
Comprehensively compare the requirements documents in docs/ against the actual
implementation (src/ and apple-fm-helper/main.swift), then bring the
derived docs back in sync. Generate a report with recommendations and questions.
Read all requirements documents in docs/: 1-overview.md,
2-architecture.md, 3-requirements.md, 4-protocol.md. If new numbered
docs have been added, include them. Also read README.md and CLAUDE.md —
both enumerate the API and module layout and drift the same way.
For each FR/NFR in 3-requirements.md, verify against the code:
src/ (or the Swift helper).Verify the wire protocol three ways agree: docs/4-protocol.md, the Node
side (src/protocol.ts + src/helper.ts), and the Swift helper
(apple-fm-helper/main.swift). The event vocabulary (delta/result/error),
the probe line shape, and the error codes must match across all three. The
reference stub tests/fixtures/stub-helper.js must also still match.
Check CLAUDE.md completeness:
docs/ appears in the docs list.src/.src/index.ts.npm run * script in package.json.vitest.config.ts.Synchronize docs/ai/code-summary.md: confirm the directory tree matches
src/ (use Glob), the Public API list matches src/index.ts, the coverage
thresholds match vitest.config.ts, and the "where do I look" entries still
resolve. Make the edits in place.
Synchronize docs/ai/requirements-summary.md: confirm each FR/NFR line
matches its row in 3-requirements.md (same status). Make the edits in place.
Final consistency pass: CLAUDE.md, README.md, and the two AI summaries
must agree with the source docs and code. Resolve any disagreement in favor of
the code / src/index.ts (API) / 3-requirements.md (behavior) /
vitest.config.ts (thresholds). The most common drift here is a new public
export or a new src/*.ts file landing without the AI summaries and CLAUDE.md
being updated — look for that explicitly.
For each: Requirement (doc + ID), Implementation (file:line), Type
(missing | different | undocumented | stale | status-wrong),
Recommendation (fix doc or fix code).
List docs, source files, exports, scripts, or thresholds that are present in code but missing from CLAUDE.md / the AI summaries (or vice versa).
The summary/doc files you updated and why (or "no changes needed").
Ambiguous requirements where the implementation made a judgment call.
No reviews yet. Tell others what you thought.
No login needed. Share how it worked for you, ask a question, or note where it did not run.
No comments yet. Be the first to write one.
Run all available tests and linters, check for anti-patterns, and generate a comprehensive code quality report
The English summary is being prepared. Showing the original description.
Check code for standardization, readability, maintenance complexity, and defensive coding practices
The English summary is being prepared. Showing the original description.
Read the latest Glassbox code review and apply all feedback annotations
The English summary is being prepared. Showing the original description.
Read the Hot Sheet worklist and work through the current priority items
The English summary is being prepared. Showing the original description.
Run as a distributed worker — continuously claim, work, and release Up Next tickets
The English summary is being prepared. Showing the original description.
Create a new bug ticket in Hot Sheet
The English summary is being prepared. Showing the original description.