원클릭으로 Manus에서 모든 스킬 실행

code-scan

스타0

포크0

업데이트2026년 2월 17일 15:59

Read-only quality scan of components. Reports problems without making changes. Uses software-base + domain profile skills.

설치

Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.

Manus에서 실행

출처

Objective-Arts

Objective-Arts/lens-dist

GitHub 저장소 열기 Creator 저장소 보기

다운로드

Manus에서 실행

/code-scan

Read-only quality analysis of target code. Identifies problems and improvement opportunities without making any changes.

Target

If a path argument is provided, scan that file/directory. If no argument, scan the code most recently discussed. Multiple paths can be provided to scan a set of components.

Step 1: Load Expert Context (MANDATORY)

Before scanning, read ALL available expert skills to build your quality lens:

Software Base (always load):

Read: .claude/skills/clarity/SKILL.md        (clear, simple code)
Read: .claude/skills/simplicity/SKILL.md     (small interfaces, composition)
Read: .claude/skills/correctness/SKILL.md    (formal correctness)
Read: .claude/skills/data-first/SKILL.md     (data structures first)
Read: .claude/skills/pragmatism/SKILL.md     (get it working)
Read: .claude/skills/security-mindset/SKILL.md (think like attacker)
Read: .claude/skills/owasp/SKILL.md          (vulnerability patterns)

Domain Profile (load what exists):

Read: .claude/skills/typescript/SKILL.md     (if .ts files)
Read: .claude/skills/react-state/SKILL.md    (if React)
Read: .claude/skills/angular-core/SKILL.md   (if Angular)
Read: .claude/skills/java/SKILL.md           (if Java)
Read: .claude/skills/python-idioms/SKILL.md  (if Python)

Skip files that don't exist. Apply all loaded principles during analysis.

Step 2: Read Target Code

Read ALL files in the target scope. Do not skim - read completely.

Step 3: Analyze (DO NOT EDIT)

Evaluate code against these criteria:

Structure

Clarity

Names reveal intent (no abbreviations except standard: API, HTTP, JSON)
No comments explaining WHAT (code should be self-documenting)
WHY comments where non-obvious
Consistent patterns throughout

Data Design

Data structures match domain
No primitive obsession (use value objects)
Illegal states unrepresentable
Immutable by default

Error Handling

All error paths explicit
No swallowed exceptions
Fail fast with clear messages
No null/undefined leaks

Security (if applicable)

Input validation at boundaries
No secrets in code
Parameterized queries (no string concat)
Output encoding

Framework Idioms (based on loaded profile)

Follows framework conventions
Uses framework patterns correctly
No anti-patterns for this framework

Output Format

## Code Scan: [target]

### Summary

| Metric | Value | Status |
|--------|-------|--------|
| Files scanned | N | - |
| Total lines | N | - |
| Largest file | N lines | ⚠️/✓ |
| Largest function | N lines | ⚠️/✓ |
| Critical issues | N | 🔴/✓ |
| Warnings | N | ⚠️/✓ |

### Critical Issues 🔴

Issues that should be fixed:

1. **[file:line]** — [description]
   - Problem: [what's wrong]
   - Impact: [why it matters]
   - Suggestion: [how to fix]

2. ...

### Warnings ⚠️

Issues to consider:

1. **[file:line]** — [description]
   - Concern: [what's questionable]
   - Consider: [improvement option]

2. ...

### Observations 💡

Not problems, but opportunities:

1. **[file]** — [observation]

### Skills Applied

- clarity: [findings summary]
- security-mindset: [findings summary]
- [domain-skill]: [findings summary]

### Files Analyzed

| File | Lines | Functions | Issues |
|------|-------|-----------|--------|
| path/to/file.ts | 245 | 12 | 2 🔴, 1 ⚠️ |
| ... | ... | ... | ... |