| name | review |
| description | Review local code changes for style, test quality, and architecture. This skill should be used when the user asks to "review code", "review my changes", "check code quality", "review the diff", or wants a pre-commit code review. Not for GitHub PR review — use /pr-review instead. |
| argument-hint | ["--staged | path | (empty=latest commit)"] |
| disable-model-invocation | true |
| allowed-tools | Read, Grep, Glob, Bash(git diff:*), Bash(git show:*), Skill |
/review — Local Code Review Workflow
Skills Rubric
| Skill | When to load |
|---|
| principles | Always — check naming, types, SOLID, function length |
| testing | Always — check AAA structure, coverage, test smells |
Step 1: Collect Change Scope
Review target: $ARGUMENTS
Determine what to review based on the argument:
--staged → git diff --stagedpath → git diff HEAD -- {path}- (empty) →
git show HEAD
Read the changed files in full to understand context — not just the diff.
Step 2: Load Skills
Load both skills. Both rubrics apply to every review.
Step 3: Style Consistency
Compare changed code against the surrounding unchanged code.
- Does naming follow the convention table in
principles?
- Is import order correct?
- Are function bodies within the length limits?
- Any
any types or magic numbers introduced?
Flag inconsistencies, not personal preferences.
Step 4: Test Quality
Evaluate the tests added or modified.
- Do tests follow AAA structure?
- Is there one logical assertion per test?
- Are test names readable as sentences?
- Any test smells from the
testing smell table?
- Does new behavior have an integration test, not just unit tests?
Step 5: Architecture Check
Scan for boundary violations in the changed code.
- Does any import cross subsystem boundaries?
- Does dependency direction follow the layer rules (Pages → Layouts → Components → Lib)?
- Is business logic leaking into the wrong layer?
Step 6: Output — Categorized Issue List
Report findings in three categories:
High — Correctness or security issues; must fix before merging.
- Examples: logic errors, missing null guards, security vulnerabilities, broken types, boundary violations
Medium — Code quality issues; should fix soon.
- Examples: naming violations, missing tests, function too long, magic numbers
Low — Style suggestions; fix if convenient.
- Examples: import order, minor naming improvements, comment clarity
Format each issue as:
[HIGH] file.ts:42 — description of the problem
[MED] file.ts:87 — description of the problem
[LOW] file.ts:12 — description of the problem
If no issues are found in a category, omit that section.
Next: run /refactor if Medium issues reveal structural problems worth cleaning up.
