claude-mods/commands/review.md

---

description: "Code review staged changes or specific files. Analyzes for bugs, style issues, security concerns, and suggests improvements."

Review - AI Code Review

Perform a comprehensive code review on staged changes or specific files.

Arguments

$ARGUMENTS

• No args: Review staged changes (git diff --cached)
• File path: Review specific file
• Directory: Review all files in directory
• --all: Review all uncommitted changes

What This Command Does

1. Identify Target Code

   • Staged changes (default)
   • Specific files/directories
   • All uncommitted changes

2. Analyze For

   • Bugs and logic errors
   • Security vulnerabilities
   • Performance issues
   • Style/convention violations
   • Missing error handling
   • Code smells

3. Provide Feedback

   • Issue severity (critical, warning, suggestion)
   • Line-specific comments
   • Suggested fixes
   • Overall assessment

Execution Steps

Step 1: Determine Scope

# Default: staged changes
git diff --cached --name-only

# If no staged changes, prompt user
git status --short

Step 2: Get Diff Content

# For staged changes
git diff --cached

# For specific file
git diff HEAD -- <file>

# For all changes
git diff HEAD

Step 3: Analyze Code

For each changed file, analyze:

Bugs & Logic

• Null/undefined checks
• Off-by-one errors
• Race conditions
• Unhandled edge cases

Security

• SQL injection
• XSS vulnerabilities
• Hardcoded secrets
• Insecure dependencies

Performance

• N+1 queries
• Unnecessary re-renders
• Memory leaks
• Blocking operations

Style

• Naming conventions
• Code organization
• Documentation gaps
• Dead code

Step 4: Format Output

# Code Review: <scope>

## Summary
- Files reviewed: N
- Issues found: X (Y critical, Z warnings)

## Critical Issues 🔴

### <filename>:<line>
**Issue**: <description>
**Risk**: <what could go wrong>
**Fix**:
\`\`\`diff
- <old code>
+ <suggested fix>
\`\`\`

## Warnings 🟡

### <filename>:<line>
**Issue**: <description>
**Suggestion**: <how to improve>

## Suggestions 🔵

### <filename>:<line>
**Suggestion**: <minor improvement>

## Overall Assessment

<1-2 sentence summary>

**Ready to commit?** Yes/No - <reasoning>

Usage Examples

# Review staged changes
/review

# Review specific file
/review src/auth/login.ts

# Review directory
/review src/components/

# Review all uncommitted changes
/review --all

# Review with specific focus
/review --security
/review --performance

Focus Flags

Flag	Focus Area
--security	Security vulnerabilities only
--performance	Performance issues only
--style	Style and conventions only
--bugs	Logic errors and bugs only
--all-checks	Everything (default)

Severity Levels

Level	Meaning	Action
🔴 Critical	Must fix before merge	Blocking
🟡 Warning	Should address	Recommended
🔵 Suggestion	Nice to have	Optional

Framework-Specific Checks

React/Next.js

• Hook rules violations
• Missing dependencies in useEffect
• Key prop issues in lists
• Server/client component boundaries

TypeScript

• any type usage
• Missing type annotations
• Incorrect generic constraints
• Type assertion abuse

Node.js

• Unhandled promise rejections
• Sync operations in async context
• Memory leak patterns
• Insecure eval/exec usage

Python

• Mutable default arguments
• Bare except clauses
• Resource leaks
• SQL string formatting

Integration

Works well with:

• /test - Generate tests for flagged issues
• /explain - Deep dive into complex code
• /checkpoint - Save state before fixing issues

Notes

• Reviews are suggestions, not absolute rules
• Context matters - some "issues" may be intentional
• Use --verbose for detailed explanations
• Reviews don't modify code - you decide what to fix