| name | review |
| description | Delegate the latest implementation diff to the code-reviewer sub-agent for a PASS/CONCERNS verdict against CLAUDE.md rules, then write the verdict to reports/<feature>-review.md. |
| disable-model-invocation | true |
/review โ Sub-Agent Code Review
Usage: /review (run after /implement and before committing)
Delegates the implementation diff to the code-reviewer sub-agent, which checks it against CLAUDE.md rules (naming, patterns, CSV escaping, JWT auth, route ordering) using the codebase-search MCP tools for structural verification. Writes the verdict to reports/<feature>-review.md.
Process
1. Identify the implementation report
Find the most recent file in reports/ matching *-implementation-report.md. That report names the feature slug and the files changed.
2. Gather the diff
Run:
git diff HEAD
If the changes are already committed (e.g., after /validate), run:
git show HEAD
3. Delegate to the code-reviewer sub-agent
Invoke the code-reviewer sub-agent with:
- The full diff text
- The feature slug (from the implementation report)
- A reminder to use
find_references / where_is to verify call sites
4. Write the review report
Write the sub-agent's verdict to reports/<feature-slug>-review.md using the Write tool. Do not modify the verdict.
5. Report back
Output:
- Path:
reports/<feature-slug>-review.md
- Verdict: PASS or CONCERNS
- If CONCERNS: list the blocking items that must be resolved before committing
Handoff: If PASS, commit. If CONCERNS, fix the flagged issues and re-run /validate then /review.
