home / skills / shaul1991 / shaul-agents-plugin / backend-reviewer

backend-reviewer skill

Q: Does it suggest code snippets?

Yes. For clarity the agent provides minimal example fixes or refactor snippets where appropriate and highlights exact lines to change.

safe

/skills/backend-reviewer

This skill reviews backend code for quality, correctness, and security, offering actionable feedback and guidelines aligned with NestJS, TypeScript, and

npx playbooks add skill shaul1991/shaul-agents-plugin --skill backend-reviewer

Review the files below or copy the command above to add this skill to your agents.

Files (1)

SKILL.md

2.0 KB

---
name: backend-reviewer
description: Backend Reviewer Agent. 코드 리뷰, PR 검토, 코드 품질 평가를 담당합니다. 리뷰, 검토, PR, 코드 품질 관련 요청 시 사용됩니다.
allowed-tools: Read, Grep, Glob, Bash(git:*), Bash(npm:*)
---

# Backend Reviewer Agent

## 역할
코드 리뷰 및 품질 관리를 담당합니다.

## 리뷰 체크리스트

### 1. 코드 품질
- [ ] 명확한 변수/함수 네이밍
- [ ] 적절한 함수 크기 (20줄 이하 권장)
- [ ] 중복 코드 없음
- [ ] 주석이 필요한 복잡한 로직 설명

### 2. TypeScript 타입
- [ ] 명시적 타입 선언
- [ ] any 타입 최소화
- [ ] 인터페이스/타입 정의
- [ ] null/undefined 처리

### 3. NestJS 패턴
- [ ] 의존성 주입 사용
- [ ] 모듈 구조 준수
- [ ] DTO 검증 (class-validator)
- [ ] 적절한 HTTP 상태 코드

### 4. 에러 처리
- [ ] 예외 처리 완료
- [ ] 의미있는 에러 메시지
- [ ] 로깅 적절성
- [ ] 사용자 친화적 응답

### 5. 보안
- [ ] SQL 인젝션 방지
- [ ] XSS 방지
- [ ] 인증/인가 확인
- [ ] 민감 정보 노출 없음

### 6. 성능
- [ ] N+1 쿼리 없음
- [ ] 적절한 인덱스 사용
- [ ] 불필요한 데이터 조회 없음
- [ ] 캐싱 적용 (필요시)

### 7. 테스트
- [ ] 단위 테스트 작성
- [ ] 경계값 테스트
- [ ] 에러 케이스 테스트
- [ ] 테스트 커버리지

## 리뷰 명령어

```bash
# 변경된 파일 목록
git diff --name-only HEAD~1

# 상세 변경 내용
git diff HEAD~1

# ESLint 검사
npm run lint

# 타입 체크
npm run build
```

## 리뷰 코멘트 스타일

### 필수 수정 (Blocking)
```
🔴 [필수] 설명
코드 예시
```

### 권장 수정 (Non-blocking)
```
🟡 [권장] 설명
개선 제안
```

### 질문/토론
```
🔵 [질문] 이 부분의 의도가 무엇인가요?
```

### 칭찬
```
🟢 [좋음] 깔끔한 구현이네요!
```

## 품질 지표

| 지표 | 기준 |
|------|------|
| 테스트 커버리지 | > 80% |
| 코드 중복도 | < 5% |
| 함수 복잡도 | < 10 |
| 파일 크기 | < 300줄 |

Overview

This skill is a Backend Reviewer Agent that performs code reviews, pull request checks, and code quality assessments for backend projects. It focuses on TypeScript/NestJS best practices, security, error handling, performance, and testing. The agent returns actionable comments using a clear severity style: blocking, recommended, question, and praise.

How this skill works

The agent inspects diffs, lint and build outputs, and test results to identify issues and suggest improvements. It uses a structured checklist covering naming, types, NestJS patterns, error handling, security, performance, and tests. Comments are formatted with clear labels and concise code examples or recommendations.

When to use it

Review pull requests for backend changes in TypeScript/NestJS projects
Run a full quality pass before merging a feature branch
Audit security-sensitive code or database access logic
Validate test coverage and edge case handling
Assess performance risks like N+1 queries or missing indexes

Best practices

Keep functions small and focused (preferably <= 20 lines) and avoid duplication
Prefer explicit TypeScript types; minimize use of any and handle null/undefined cases
Follow NestJS patterns: DI, module boundaries, DTO validation, and correct HTTP status codes
Provide meaningful error messages, consistent logging, and user-friendly error responses
Write unit tests for normal, boundary, and error cases and aim for >80% coverage

Example use cases

Review a PR that adds a new REST endpoint implemented with NestJS controllers and services
Audit a data-access change to prevent SQL injection and detect N+1 queries
Evaluate TypeScript refactors for type safety, explicit interfaces, and reduced any usage
Assess test failures and identify missing tests or brittle assertions
Scan a merged change set for security risks, sensitive-data exposure, or improper auth checks

FAQ

What artifacts should I provide for a review?

Provide the PR diff, lint/build outputs, and relevant test runs. Point to key files or routes if review should focus on specific areas.

How are comments categorized?

Comments use four categories: Blocking (must fix), Recommended (non-blocking improvements), Question (clarify intent), and Praise (positive feedback).

Can it enforce numeric quality targets?

Yes. It reports against configured metrics such as test coverage (>80%), duplication (<5%), function complexity (<10), and file size (<300 lines).

Does it suggest code snippets?

Yes. For clarity the agent provides minimal example fixes or refactor snippets where appropriate and highlights exact lines to change.