Deprecated in v5.2 — 이 커맨드는 /deep-work auto-flow에서 자동 실행됩니다. 수동 호출도 여전히 가능합니다. 통합 워크플로우는 /deep-work을 참고하세요.

Code Insight Analysis

You are performing a Code Insight Analysis — measuring code metrics, complexity indicators, and dependency patterns to provide informational reports. This is the Insight tier of the 3-tier Quality Gate system.

Critical Constraints

• DO NOT modify any code files. This is an analysis-only operation.
• Read, analyze, and report findings.
• Insight results NEVER block the workflow. They are purely informational.
• Save analysis results to file when in workflow mode.

Instructions

1. Determine operating mode

Resolve the current session's state file:

1. If DEEP_WORK_SESSION_ID env var is set → .claude/deep-work.${DEEP_WORK_SESSION_ID}.md
2. If .claude/deep-work-current-session pointer file exists → read session ID → .claude/deep-work.${SESSION_ID}.md
3. Legacy fallback → .claude/deep-work.local.md

Set $STATE_FILE to the resolved path.

Check if $STATE_FILE exists and has an active session (current_phase is not idle and not empty).

Workflow Mode (active deep-work session):

• Read work_dir from the state file
• Set WORK_DIR to the value of work_dir
• Read $WORK_DIR/plan.md to extract the list of files (from "Files to Modify" section)
• Cross-reference with actual changes: git diff --name-only from baseline
• If $WORK_DIR/file-changes.log exists, use it for additional file tracking data
• Analysis scope: files listed in plan.md that were actually modified

Standalone Mode (no active session):

• If $ARGUMENTS is provided: use as target (file path, directory, or glob pattern)
• If $ARGUMENTS is empty: detect scope automatically:
  1. Check git diff --name-only HEAD~1 for recently changed files
  2. If not a git repo or no changes, use current directory
• Analysis scope: all code files in detected scope (exclude node_modules, .git, pycache, build, dist, .next, vendor, etc.)

2. Collect analysis targets

Gather the list of files to analyze. For each file:

• Skip files that are clearly not code (README.md, .json config, .env, .lock, etc.)
• Skip files smaller than 3 lines (trivial)
• Skip auto-generated files (migrations, lock files, bundled output, .min.js)

If the total number of files exceeds 50, prioritize:

1. Files with the most lines of code
2. Files modified most recently
3. Files explicitly listed in plan.md (workflow mode)

Display progress:

ℹ️ Insight 분석 대상: [N]개 파일
  - src/auth/service.ts (245 lines)
  - src/models/user.ts (180 lines)
  - ...

If exceeding 50 files:

⚠️ 분석 대상이 50개를 초과합니다. 상위 50개 파일만 분석합니다.

3. Run built-in analyses

Execute 4 categories of analysis. Each analysis should be resilient — if one fails, continue with the rest.

3A. File Metrics

For each target file, measure:

• Lines of code (excluding blank lines and comment-only lines)
• Function/method count: Use grep -c patterns appropriate for the language:
  • JS/TS: function , => {, method definitions in classes
  • Python: def
  • Go: func
  • Rust: fn
  • Java/Kotlin: method patterns
  • Other: best-effort grep
• Export count (JS/TS: export , Python: __all__, Go: uppercase functions)

Generate summary table:

## A. 파일 메트릭

| 파일 | 코드 줄 수 | 함수 수 | Export 수 |
|------|-----------|---------|----------|
| src/auth.ts | 245 | 12 | 5 |
| src/db.ts | 180 | 8 | 3 |

**합계**: 코드 425줄, 함수 20개, Export 8개

3B. Complexity Indicators

For each target file, check:

• Long files: files exceeding 300 lines → flag with line count
• Long functions: functions exceeding 50 lines → identify function name and line range
  • Use heuristic: find function declarations, then count lines until matching closing brace/dedent
• Deep nesting: maximum indentation depth (count leading spaces/tabs)
  • Threshold: > 4 levels of nesting

Generate findings:

## B. 복잡도 지표

### 대형 파일 (300줄 초과)
| 파일 | 줄 수 | 상태 |
|------|------|------|
| src/auth.ts | 456 | ⚠️ 대형 |

### 장함수 (50줄 초과)
| 파일 | 함수명 | 줄 수 |
|------|--------|------|
| src/auth.ts | handleLogin | 82 |

### 깊은 중첩 (4단계 초과)
| 파일 | 최대 깊이 | 위치 |
|------|----------|------|
| src/parser.ts | 6 | line 45-78 |

**요약**: 대형 파일 N개, 장함수 N개, 깊은 중첩 N개

If no issues found for a subcategory, display: "없음 ✅"

3C. Dependency Analysis

For each target file, parse import/require statements:

• JS/TS: import ... from '...', require('...')
• Python: import ..., from ... import ...
• Go: import "...", import (...)
• Other: best-effort pattern matching

Build a simple adjacency list of internal dependencies (skip external packages).

Check for:

• Circular dependencies: Simple DFS cycle detection among internal modules
  • If cycles found, list each cycle as: A → B → C → A
• Import depth: Maximum chain length in the dependency graph
• Hub files: Files imported by 5+ other files (potential coupling hotspot)

Generate findings:

## C. 의존성 분석

### Import 통계
| 파일 | Import 수 | 내부 | 외부 |
|------|----------|------|------|
| src/auth.ts | 8 | 3 | 5 |

### 순환 참조
없음 ✅
(or: ⚠️ 순환 참조 발견: src/a.ts → src/b.ts → src/a.ts)

### Hub 파일 (5+ 참조)
| 파일 | 참조 수 |
|------|---------|
| src/utils.ts | 12 |

### Import 깊이
최대 Import 체인: 4 (src/page.ts → src/auth.ts → src/db.ts → src/config.ts)

3D. Change Summary (workflow mode only)

Only execute this section when in workflow mode with an active session.

Read $WORK_DIR/file-changes.log if it exists, or fall back to git diff --stat.

Generate:

## D. 변경 요약

| 항목 | 값 |
|------|---|
| 총 변경 파일 수 | N |
| 신규 생성 | N |
| 수정 | N |
| 삭제 | N |
| 추가 줄 수 | +N |
| 삭제 줄 수 | -N |

### 파일별 수정 횟수 (file-changes.log 기준)
| 파일 | 수정 횟수 |
|------|----------|
| src/auth.ts | 5 |
| src/db.ts | 3 |

If neither file-changes.log nor git data is available, display:

변경 데이터 없음 — git diff 또는 file-changes.log를 사용할 수 없습니다.

4. Execute user-defined Insight gates (workflow mode)

If in workflow mode and $WORK_DIR/plan.md contains a Quality Gates table with ℹ️ markers:

1. Parse each ℹ️ gate: extract gate name and command
2. Execute each command
3. Record the result (output, exit code) — but NEVER treat failure as blocking
4. Include results in the Insight Gates section

## 사용자 정의 Insight Gates

| Gate | 명령어 | 결과 | 출력 |
|------|--------|------|------|
| Complexity Score | `npx complexity-report` | ℹ️ INFO | score: 12.3 |

If a user-defined Insight gate command fails:

| Gate Name | `command` | ℹ️ SKIP | 명령어 실행 실패 (exit code: N) |

5. Generate overall summary

Combine all analyses into a concise overview:

## 종합 인사이트 요약

| 카테고리 | 핵심 수치 | 주목 포인트 |
|----------|----------|------------|
| 파일 메트릭 | N파일, N줄, N함수 | — |
| 복잡도 | 대형 N / 장함수 N / 깊은 중첩 N | [가장 심각한 항목] |
| 의존성 | 순환 참조 N / Hub N / 최대 깊이 N | [가장 심각한 항목] |
| 변경 요약 | +N/-N줄, N파일 | — |

**판정**: ℹ️ Insight — 워크플로우 차단 없음, 참고용 정보 제공

6. Save results

Workflow Mode:

• Write full results to $WORK_DIR/insight-report.md
• Display summary in terminal:

  ℹ️ Insight 분석 완료 — 결과가 $WORK_DIR/insight-report.md에 저장되었습니다.
  

Standalone Mode:

• Display full results in terminal
• Ask user: "분석 결과를 파일로 저장할까요? (기본: 아니오)"
• If yes, save to ./insight-report.md

7. Workflow integration (workflow mode only)

If called as a Quality Gate during Test phase:

• Record results in quality-gates.md as Insight entry
• Results do NOT affect pass/fail determination — always informational only

If called outside the Test phase:

• Still run the analysis
• Note: "deep-work 워크플로우 활성 — 결과가 $WORK_DIR/insight-report.md에 저장됩니다"

8. Error resilience

If any analysis section fails (e.g., no git repo, unsupported language):

• Log the error inline: ⚠️ [Section Name] 분석 실패: [reason]
• Continue with remaining analyses
• NEVER abort the entire analysis due to a single section failure
• NEVER return a non-zero/blocking result