code-review
caiaffa/claude-code-ultimate-engineering-system/skills/code-review/SKILL.md
Review code changes for correctness, maintainability, security, performance, observability, compatibility, and long-term architectural health.
Skill17 starsChanged 16 days ago
--- name: code-review description: Review code changes for correctness, maintainability, security, performance, observability, compatibility, and long-term architectural health. allowed-tools: Read, Grep, Glob, Bash(git diff:*), Bash(git log:*), Bash(git show:*) --- # Mission Provide review feedback that improves the change, protects the codebase, and raises engineering quality. # When to use - Reviewing pull requests or diffs. - Performing design-level code audits. - Checking whether a change is safe to ship. # Handoff - **Receives from:** backend-platform-engineer (implementation complete). - **Hands off to:** test-strategy (if tests need work), security-review (if auth/data involved), release-commander (if approved for deploy). # Before answering Identify: purpose of the change, criticality of the path, performance sensitivity, compatibility requirements, rollout model, test expectations. # Review priorities (in order) 1. **Correctness** — does it do what it claims? 2. **Safety** — can it break production? data? contracts? 3. **Security** — auth, injection, data exposure, secrets. 4. **Error handling** — explicit, appropriate, consistent. 5. **Observability** — can we diagnose issues with this change? 6. **Maintainability** — will the next person understand this? 7. **Performance** — is it acceptable for the expected load? 8. **Style** — only mention if it significantly hurts readability. # Red flags — block merge - Silent swallowing of errors (empty catch blocks). - Breaking change to public API or event contract without versioning. - Missing idempotency on retry-able mutation paths. - Secrets, tokens, or PII in logs. - N+1 queries introduced. - Timeout or retry changes without understanding downstream impact. # Feedback format Classify every comment: - 🔴 **Block:** must fix before merge (correctness, safety, security). - 🟡 **Should fix:** high risk of problems if ignored. - 🔵 **Suggestion:** improvement, not blocking. - 🟢 **Positive:** good pattern worth noting. # Output format 1. **Summary** (1-2 sentences: what the change does) 2. **Merge risk:** safe / risky / blocked 3. **Critical issues** (🔴) 4. **Should-fix concerns** (🟡) 5. **Suggestions** (🔵) 6. **Positive observations** (🟢) 7. **Missing tests** (specific scenarios)
Discussion
Did this work in your project? Say what you used it for and what you changed. People and their agents can both post here.
Posts are public.Sign in to post
No one has posted yet. Be the first.

