This Claude skill automates the code review process by analyzing source code against professional quality assurance guidelines. It identifies security vulnerabilities, logic bugs, and performance bottlenecks while providing actionable feedback and structured reports to ensure high-quality, maintainable software delivery.

When should I use review?

review is useful in the following scenarios: • Pre-commit Quality Checks: Use the skill to analyze staged changes (git diff) before committing to ensure no security leaks or obvious bugs are introduced. • Automated PR Reviews: Streamline the pull request process by getting an immediate, structured evaluation of code changes, including priority-based issue categorization. • Security and Performance Auditing: Specifically scan files for critical vulnerabilities like SQL injection or XSS, and identify performance issues like N+1 queries or memory leaks. • Post-Generation Verification: Automatically trigger a review after Claude generates code to verify that the output meets team-specific standards and best practices. • Refactoring Support: Get actionable suggestions and 'Good vs. Bad' code examples to improve the maintainability and readability of legacy codebases.

name	review
description	Performs code review following team's quality assurance guidelines. Use when the user asks to "review code", "check this code", "리뷰해줘", "코드 리뷰", "코드 검토", "이 코드 봐줘", "이거 괜찮아?", "확인해줘", "문제 없어?", or after writing code when they want feedback on the changes just made.
allowed-tools	Read, Glob, Grep, Bash(git diff:), Bash(git log:), Bash(git show:*)
version	1.0.0

Document: SKILL.md Role: Code Review Assistant Priority: High - Quality assurance automation Applies To: Code review in any project Optimized For: Claude 4.5 (Sonnet/Opus) Last Updated: 2025-12-28 This skill is auto-discovered by Claude when users request code review tasks. It can also be triggered naturally after code has been written, when the user wants to verify the quality of recent changes.

Code Review Skill

This skill performs code reviews following the team's quality assurance guidelines.

Source of Truth

Review Format: output-formats.md
Quality Standards: quality-assurance.md
Security Guidelines: security.md

When to Activate

This skill activates in these scenarios:

Explicit request: User asks for code review directly
After code writing: User asks "이거 괜찮아?", "확인해줘", "문제 없어?" after Claude wrote code
Change verification: User wants to check staged/unstaged changes
PR preparation: Before committing, user wants quality check

Review Principles

When reviewing code, follow these core principles:

Actionable feedback: Every issue must have a clear fix suggestion
Prioritized issues: Critical issues first, then improvements
Balanced perspective: Acknowledge good practices, not just problems
Educational tone: Explain WHY something is an issue

Instructions

Step 1: Understand the Scope

Identify what code to review:
- Just written code: Review the changes Claude just made in this conversation
- Specific file(s): Files provided by user
- Recent changes: git diff for unstaged, git diff --cached for staged
- PR/commit changes: git show <commit>
Read the code thoroughly before commenting

Step 2: Analyze Code

Check for issues in these categories:

Category	Priority	Examples
Security	🔴 Critical	SQL injection, XSS, secrets in code
Bugs	🔴 Critical	Logic errors, null references, race conditions
Performance	🟡 Medium	N+1 queries, unnecessary loops, memory leaks
Maintainability	🟡 Medium	Long functions, unclear names, missing docs
Style	🟢 Low	Formatting, conventions, minor improvements

Step 3: Apply Self-Review Checklist

Before providing feedback, verify against this checklist:

All tests pass (if applicable)
Edge cases handled
Performance impact considered
No security vulnerabilities
Error messages are user-friendly
No commented-out code
No debug statements (console.log, print, etc.)

Step 4: Format Review Output

Use this structured format for review results:

## 코드 리뷰 결과

### 📊 전체 평가
- **품질**: [상/중/하]
- **주요 이슈**: [N개 발견]
- **긴급도**: [즉시 수정 필요/개선 권장/양호]

### 🔴 Critical Issues (우선순위: 높음)

**[Issue Title]** (`file:line`)
- **Problem**: [명확한 문제 설명]
- **Impact**: [영향 범위와 위험도]
- **Fix**: [구체적인 수정 방법]

\`\`\`[language]
// Bad
[problematic code]

// Good
[fixed code]
\`\`\`

### 🟡 Improvements (우선순위: 중간)

1. **[Issue Title]** (`file:line`)
   - [문제 설명]
   - [개선 방법]

### 🟢 Good Practices

- ✅ [잘 작성된 부분 1]
- ✅ [잘 작성된 부분 2]

### ✅ Action Items

1. [ ] [우선순위별 작업 목록]
2. [ ] [...]

Simplified Output (for small changes)

When reviewing small changes or code just written, use a lighter format:

## 리뷰 결과

✅ **양호** - 주요 문제 없음

### 확인 사항
- ✅ [확인된 항목 1]
- ✅ [확인된 항목 2]

### 개선 제안 (선택)
- 💡 [사소한 개선 사항]

Issue Templates

Security Issue

**[보안 취약점 유형]** (`file:line`)
- **Problem**: [취약점 설명]
- **Impact**: [공격 시나리오와 피해 범위]
- **Fix**: [수정 방법]
- **Reference**: [OWASP 또는 관련 문서 링크]

Performance Issue

**[성능 문제 유형]** (`file:line`)
- **Problem**: [현재 성능 문제]
- **Impact**: [예상 성능 저하]
- **Fix**: [최적화 방법]

Maintainability Issue

**[유지보수 문제 유형]** (`file:line`)
- **Problem**: [현재 코드의 문제점]
- **Impact**: [향후 유지보수 어려움]
- **Fix**: [리팩토링 제안]

Quality Ratings

품질 평가 기준

Rating	Criteria
상 (High)	No critical issues, minor improvements only, follows best practices
중 (Medium)	No critical issues, some improvements needed, mostly follows conventions
하 (Low)	Critical issues found, significant refactoring needed

긴급도 평가 기준

Urgency	Criteria
즉시 수정 필요	Security vulnerabilities, data loss risk, production bugs
개선 권장	Performance issues, code smells, missing tests
양호	Only minor style/formatting suggestions

Response Language

Review comments: Korean (한국어)
Code examples: English (comments, variable names)
Technical terms: Keep in English (e.g., SQL injection, N+1, refactoring)

review

When & Why to Use This Skill

Use Cases