Agent skill

review-changes

Review uncommitted code changes for bugs, security issues, and style violations.

Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/review-changes-andersonsrepo-ai-harness

SKILL.md

Code Review

Review uncommitted code changes for bugs, security issues, and style violations.

Changes to Review

!git diff --cached 2>/dev/null || echo "(nothing staged)" !git diff 2>/dev/null || echo "(no unstaged changes)" !git ls-files --others --exclude-standard 2>/dev/null || echo "(no untracked files)"

Scope

Parse $ARGUMENTS to determine what to review:

  • --staged — only staged changes (git diff --cached)
  • --all — all changes including unstaged and untracked
  • <file-path> — review a specific file
  • No args — review staged changes if any, otherwise all changes

Review Checklist

For each changed file, evaluate:

Correctness

  • Logic errors or off-by-one mistakes
  • Missing null/undefined checks at system boundaries
  • Incorrect error handling (swallowed errors, wrong catch scope)
  • Race conditions in async code
  • Resource leaks (unclosed files, connections, watchers)

Security

  • Command injection (unsanitized input in shell commands)
  • Path traversal (user input in file paths)
  • Secrets or credentials in code or config
  • Exposed env vars or API keys
  • SQL injection (if applicable)

Style & Conventions

  • TypeScript types (no unnecessary any)
  • Consistent naming (camelCase functions, PascalCase types)
  • Import organization
  • Dead code or unused variables

AI Harness Specific

  • Claude CLI spawning follows the 7 rules in CLAUDE.md
  • -- separator before prompt arguments
  • CLAUDE* env vars stripped in subprocess spawning
  • Disallowed tools guardrails included in Claude invocations
  • File-based output pattern used (not pipes)

Output Format

## Code Review: <scope>

### Summary
<1-2 sentence overview>

### Issues Found
1. **[SEVERITY]** file.ts:L42 — Description of issue
   Suggestion: <how to fix>

2. **[SEVERITY]** file.ts:L87 — Description of issue
   Suggestion: <how to fix>

### Approved
- file.ts — No issues found

### Verdict
APPROVE / REQUEST_CHANGES / NEEDS_DISCUSSION

Severity levels: CRITICAL (must fix), WARNING (should fix), NITS (optional improvement).

Expand your agent's capabilities with these related and highly-rated skills.

Didn't find tool you were looking for?

Be as detailed as possible for better results