소스 정보
- 저장소
- tomevault-io/skills-registry
- 최근 소스 활동
- 2026년 5월 11일 15:30
- 감지된 SKILL.md 언어
- 영어
- 스타
- 0
- 포크
- 0
설치 방법
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
소스 파일 검토
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
메뉴
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
SOC 직업 분류 기준
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/tomevault-io/skills-registry --skill code-review명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
SKILL.md 표시 중
| Use when this capability is needed.
> Use when this capability is needed.
Review architecture and API design for the vfs-s3 project. Use when the user mentions @architect, asks to review an issue's design, discuss module boundaries, API shape, or architectural decisions for vfs-s3. Also trigger when the user wants to create an ADR (Architecture Decision Record) or evaluate a technical approach for the project. Intended for dispatch from Codex automation or Claude routines; GitHub trigger phrase: @vfs-s3-bot please prepare design doc Use when this capability is needed.
| name | code-review |
| description | name: arcanea-code-review Use when this capability is needed. |
name: arcanea-code-review description: Conduct thorough, constructive code reviews that improve code quality and team knowledge. Focuses on what matters - architecture, logic, security, maintainability - while avoiding bikeshedding. version: 2.0.0 author: Arcanea tags: [code-review, quality, collaboration, development] triggers:
"A review is not a judgment. It is a gift of attention that makes both the code and the coder stronger."
✓ Knowledge sharing
✓ Quality assurance
✓ Learning opportunity
✓ Documentation check
✓ Collaboration ritual
❌ Gatekeeping
❌ Proving superiority
❌ Stylistic bikeshedding
❌ Blocking progress
❌ Personal criticism
Focus effort where it matters most:
╔════════════════════════════════════════════════════════════╗
║ REVIEW PRIORITY ║
╠════════════════════════════════════════════════════════════╣
║ ║
║ 🔴 CRITICAL (Block merge) ║
║ ├── Security vulnerabilities ║
║ ├── Data loss risks ║
║ ├── Breaking changes without migration ║
║ └── Logic errors affecting correctness ║
║ ║
║ 🟠 IMPORTANT (Should fix before merge) ║
║ ├── Bugs and edge cases ║
║ ├── Performance issues ║
║ ├── Missing tests for new logic ║
║ └── Architectural concerns ║
║ ║
║ 🟡 SUGGESTIONS (Nice to have) ║
║ ├── Readability improvements ║
║ ├── Better naming ║
║ ├── Documentation additions ║
║ └── Minor refactoring opportunities ║
║ ║
║ ⚪ NITPICKS (Optional, don't block) ║
║ ├── Style preferences ║
║ ├── Formatting ║
║ └── Subjective choices ║
║ ║
╚════════════════════════════════════════════════════════════╝
□ No hardcoded secrets/credentials
□ Input validation present
□ No SQL injection risks
□ No XSS vulnerabilities
□ Authentication/authorization correct
□ Sensitive data handled properly
□ Dependencies are up to date
□ No debug/admin backdoors
□ Code does what it claims to do
□ Edge cases handled
□ Error handling is appropriate
□ Null/undefined handled safely
□ Race conditions considered
□ State management is correct
□ No obvious bugs
□ Follows project patterns
□ Separation of concerns
□ Dependencies flow correctly
□ No circular dependencies
□ Appropriate abstraction level
□ DRY (Don't Repeat Yourself)
□ SOLID principles where applicable
□ Code is readable
□ Functions are reasonably sized
□ Names are clear and accurate
□ Complex logic is commented
□ No magic numbers/strings
□ Easy to modify in future
□ No unnecessary complexity
□ Tests exist for new functionality
□ Tests cover edge cases
□ Tests are readable
□ Tests actually test something
□ No testing implementation details
□ Existing tests still pass
□ No obvious performance issues
□ Database queries are efficient
□ No N+1 query problems
□ Appropriate caching
□ Memory usage reasonable
□ No blocking operations on main thread
**Level**: [Critical/Important/Suggestion/Nitpick]
**What**: [Specific issue]
**Why**: [Impact or concern]
**How**: [Suggested fix or alternative]
🔴 **Critical: SQL Injection Risk**
**Line 45**: `db.query("SELECT * FROM users WHERE id = " + userId)`
This is vulnerable to SQL injection. An attacker could delete data
or access unauthorized information.
**Suggested fix**:
```js
db.query("SELECT * FROM users WHERE id = ?", [userId])
#### Important Suggestion
```markdown
🟠 **Important: Missing Error Handling**
**Line 78**: `const data = await fetchUser(id)`
If fetchUser throws, this will crash the request handler and
return a 500 to the user.
**Suggested fix**:
```js
try {
const data = await fetchUser(id);
} catch (error) {
logger.error('Failed to fetch user', { id, error });
return res.status(404).json({ error: 'User not found' });
}
#### Suggestion
```markdown
🟡 **Suggestion: Naming Clarity**
**Line 32**: `const d = new Date()`
Single-letter variable names reduce readability.
**Consider**: `const createdAt = new Date()`
⚪ **Nitpick** (optional, non-blocking)
**Line 15**: Would prefer `const` over `let` here since it's never reassigned.
AVOID:
- "You should..."
- "This is wrong"
- "Why would you..."
- "Obviously..."
PREFER:
- "Consider..."
- "What if we..."
- "I wonder if..."
- "One option might be..."
Questions often work better than commands:
"Could this throw if the user doesn't exist?"
vs
"This will crash when user doesn't exist!"
1. Read the PR description
2. Understand the goal
3. Scan all files changed
4. Get the big picture
1. Read each file carefully
2. Check logic flow
3. Look for bugs and issues
4. Note questions
1. Check how it fits with existing code
2. Consider future implications
3. Look for missing tests
4. Consider edge cases
If PR is too big to review effectively:
1. Request it be split into smaller PRs
2. Focus on highest-risk files first
3. Review in multiple sessions
4. Trust tests for mechanical changes
Key questions:
- Does behavior remain identical?
- Are there tests proving behavior is preserved?
- Is the new structure actually better?
- Is this the right time for this refactor?
Key questions:
- Does it actually fix the bug?
- Is there a test that would have caught this?
- Could this fix break something else?
- Is the root cause addressed?
Key questions:
- Does it meet requirements?
- Is it complete or partial?
- Are there edge cases?
- Is it testable and tested?
- Does it fit the architecture?
Key questions:
- Is this update necessary?
- Are there breaking changes?
- Have changelogs been reviewed?
- Do tests still pass?
- Any security advisories?
Before requesting review, check:
□ Code compiles/passes linter
□ Tests pass
□ Changes match PR description
□ No debug code left in
□ No commented-out code
□ No unrelated changes
□ Commit messages are clear
□ Documentation updated if needed
□ PR is reasonably sized
□ Ready for someone else to read
Review within 24 hours if possible.
Blocked authors = blocked productivity.
If you can't review, say so.
Review thoroughly the first time.
Multiple rounds of "one more thing" is frustrating.
Group all feedback in one review.
Praise what's good, not just what's wrong.
"Nice approach to this problem"
"Clean solution for the edge case"
Genuine appreciation builds trust.
Reviews aren't about winning.
The goal is better code AND better coders.
Be willing to be wrong.
Defer to author on judgment calls.
🔴 CRITICAL: - Must fix before merge
🟠 IMPORTANT: - Should fix before merge
🟡 SUGGESTION: - Improvement idea
⚪ NIT: - Take it or leave it
❓ QUESTION: - Need clarification
💭 THOUGHT: - Something to consider
👍 NICE: - Positive feedback
If you only have 10 minutes:
1. Read PR description (1 min)
2. Scan file changes (2 min)
3. Check highest-risk code (5 min)
4. Verify tests exist (2 min)
Flag if needs deeper review.
- Very large diffs (>500 lines)
- No tests for new logic
- Commented-out code
- TODOs without tickets
- Copy-pasted code blocks
- Complex nested logic
- Magic numbers/strings
- Ignored error handling
- Hardcoded values
- Console.log/print statements
"Review the code, not the coder. The goal is software we're all proud of."
Converted and distributed by TomeVault — claim your Tome and manage your conversions.