/fsdarch:analyze

현재 FSD 프로젝트 구조를 분석하고 상세 리포트를 생성합니다.

Prerequisites

• .fsd-architect.json 설정 파일 존재 (없으면 /fsdarch:init 먼저 실행)

---

EXECUTION INSTRUCTIONS

When /fsdarch:analyze is invoked, Claude MUST perform these steps in order:

1. Load configuration - Read .fsd-architect.json (error E104 if missing)
2. Check cache - Use skill: cache-manager for incremental analysis
3. Use skill: layer-detector - Scan all layers and slices
4. Use skill: boundary-checker - Analyze import dependencies
5. Calculate health score - Score based on 5 criteria
6. Generate report - Display formatted analysis results
7. Update cache - Save analysis results for incremental updates

---

Flow Overview

┌───────────────────────────────────────────────────────────────┐
│                      /fsdarch:analyze                         │
├───────────────────────────────────────────────────────────────┤
│  1. Load .fsd-architect.json                                  │
│     ├─ Not found → Error E104, suggest /fsdarch:init          │
│     └─ Found → Continue                                       │
│                                                               │
│  2. Check cache (unless --force)                              │
│     ├─ Valid cache exists → Use incremental analysis          │
│     └─ No cache / stale → Full scan                           │
│                                                               │
│  3. Invoke skill: layer-detector                              │
│     → Scan all layers, slices, segments                       │
│     → Count files per layer/slice                             │
│                                                               │
│  4. Invoke skill: boundary-checker                            │
│     → Parse import statements                                 │
│     → Build dependency graph                                  │
│     → Detect violations (E201, E202, E203, E204, E205)        │
│                                                               │
│  5. Calculate health score (100 points)                       │
│     → Layer hierarchy (30%)                                   │
│     → Public API usage (25%)                                  │
│     → Slice isolation (20%)                                   │
│     → Naming consistency (15%)                                │
│     → Segment structure (10%)                                 │
│                                                               │
│  6. Display report with issues and recommendations            │
│                                                               │
│  7. Update .fsd-architect.cache.json                          │
└───────────────────────────────────────────────────────────────┘

---

Execution Flow

Step 1: Load Configuration

Action: Read and parse .fsd-architect.json

1. Use Read tool to read .fsd-architect.json
2. If file not found:
   → Display error E104 and stop
3. If JSON parse error:
   → Display error E401 with line number
4. Validate required fields:
   → srcDir (required)
   → layers (required)
5. Extract configuration values for subsequent steps

Read tool command:

Read: .fsd-architect.json

Error output if missing:

[E104] Configuration not found

.fsd-architect.json is required for this command.

Solution: Initialize first:
  /fsdarch:init

Progress output:

> Loading configuration...

Step 1.5: Check Cache (Optional)

Action: Invoke skill: cache-manager for incremental analysis

1. Skip if --force flag is set
2. Read .fsd-architect.cache.json (if exists)
3. Compare file mtimes with cached values
4. Identify files that changed since last analysis
5. Only re-analyze changed files (incremental mode)

Cache benefits:

• Faster analysis on large projects
• Only re-scan modified files
• Preserve previous violation history

Progress output (if using cache):

> Using cached analysis (15 files changed since last run)

Step 2: Scan Layer Structure

Action: Invoke skill: layer-detector

1. For each layer in config.layers:
   → Glob for {srcDir}/{layerPath}/**/*
   → If sliced layer: list immediate subdirectories as slices
   → For each slice: list segments (ui/, model/, api/, lib/)
   → Count files per layer/slice

2. Build LayerInfo structure for each layer:
   {
     name: "features",
     path: "src/features",
     slices: [
       { name: "auth", segments: ["ui", "model", "api"], hasPublicApi: true, fileCount: 12 },
       { name: "cart", segments: ["ui", "model"], hasPublicApi: true, fileCount: 8 }
     ],
     totalFiles: 156,
     isSliced: true
   }

Glob commands to execute:

IMPORTANT: Memory Safety - Use pagination for large projects

For projects with many files, use head_limit and offset to batch file processing:

# RECOMMENDED: Batch processing with head_limit (max 1000 files per batch)
# This prevents memory exhaustion on large codebases (10,000+ files)

# For each layer defined in config (with pagination)
Glob: "{srcDir}/app/**/*.{ts,tsx,js,jsx}" head_limit=1000 offset=0
Glob: "{srcDir}/pages/**/*.{ts,tsx,js,jsx}" head_limit=1000 offset=0
Glob: "{srcDir}/widgets/**/*.{ts,tsx,js,jsx}" head_limit=1000 offset=0
Glob: "{srcDir}/features/**/*.{ts,tsx,js,jsx}" head_limit=1000 offset=0
Glob: "{srcDir}/entities/**/*.{ts,tsx,js,jsx}" head_limit=1000 offset=0
Glob: "{srcDir}/shared/**/*.{ts,tsx,js,jsx}" head_limit=1000 offset=0

# If more than 1000 files in a layer, continue with offset:
# Glob: "{srcDir}/features/**/*.{ts,tsx,js,jsx}" head_limit=1000 offset=1000
# ... repeat until all files processed

# For sliced layers - get slice list
Glob: "{srcDir}/features/*/"
Glob: "{srcDir}/entities/*/"

# Check public API for each slice
Glob: "{srcDir}/features/*/index.{ts,tsx,js,jsx}"

Batch Processing Algorithm:

const BATCH_SIZE = 1000;

async function batchGlob(pattern: string): Promise<string[]> {
  const allFiles: string[] = [];
  let offset = 0;

  while (true) {
    const batch = await glob(pattern, { head_limit: BATCH_SIZE, offset });
    if (batch.length === 0) break;
    allFiles.push(...batch);
    if (batch.length < BATCH_SIZE) break;  // Last batch
    offset += BATCH_SIZE;
  }

  return allFiles;
}

TypeScript interfaces:

interface LayerInfo {
  name: string;
  path: string;
  slices: SliceInfo[];
  totalFiles: number;
  isSliced: boolean;
}

interface SliceInfo {
  name: string;
  segments: string[];
  hasPublicApi: boolean;
  fileCount: number;
}

Progress output:

> Scanning 6 layers, 20 slices...

Step 3: Analyze Dependencies

Action: Invoke skill: boundary-checker

1. For each TypeScript/JavaScript file in the project:
   → Read file content
   → Parse import/require statements
   → Extract import paths

2. For each import:
   → Resolve to actual file path
   → Determine source layer/slice
   → Determine target layer/slice
   → Check against FSD rules

3. Build dependency graph:
   → Track all layer → layer dependencies
   → Track all slice → slice dependencies
   → Identify violations

Grep commands to find imports:

# Find all import statements
Grep: "^import .* from ['\"]" --type ts
Grep: "require\\(['\"]" --type ts

# Focus on potential violations (cross-layer/slice imports)
Grep: "from ['\"]@features/" --type ts
Grep: "from ['\"]@entities/" --type ts
Grep: "from ['\"]\\.\\./\\.\\." --type ts  # Relative cross-boundary

Violation detection rules:

// Layer hierarchy (lowest = 1, highest = 6)
const LAYER_ORDER = {
  shared: 1,
  entities: 2,
  features: 3,
  widgets: 4,
  pages: 5,
  app: 6
};

// E203: Higher layer import from lower layer
function checkLayerViolation(source: Layer, target: Layer): boolean {
  return LAYER_ORDER[source] < LAYER_ORDER[target];
}

// E201: Cross-slice import (same layer)
function checkCrossSlice(source: Slice, target: Slice, layer: Layer): boolean {
  return source.layer === target.layer && source.name !== target.name;
}

// E202: Public API sidestep
function checkPublicApiSidestep(importPath: string): boolean {
  // Import should be @entities/user, not @entities/user/model/types
  return importPath.split('/').length > 2 && !importPath.endsWith('/index');
}

// E204: Missing public API
function checkMissingPublicApi(slice: Slice): boolean {
  return !slice.hasPublicApi;
}

Progress output:

> Analyzing dependencies...

Step 4: Calculate Health Score

Action: Compute weighted health score (100 points max)

Scoring criteria:

Criterion	Weight	How to Calculate
Layer Hierarchy	30%	100 - (E203 violations * 10), min 0
Public API Usage	25%	(slices with index.ts / total slices) * 100 - (E202 violations * 5)
Slice Isolation	20%	100 - (E201 violations * 10), min 0
Naming Consistency	15%	(slices matching dominant pattern / total slices) * 100
Segment Structure	10%	(slices with standard segments / total slices) * 100

Calculation algorithm:

function calculateHealthScore(analysis: AnalysisResult): number {
  const weights = {
    layerHierarchy: 0.30,
    publicApi: 0.25,
    sliceIsolation: 0.20,
    namingConsistency: 0.15,
    segmentStructure: 0.10
  };

  // Layer Hierarchy: Penalize E203 violations
  const e203Count = analysis.violations.filter(v => v.code === 'E203').length;
  const layerScore = Math.max(0, 100 - e203Count * 10);

  // Public API: Check index.ts presence, penalize E202
  const slicesWithApi = analysis.slices.filter(s => s.hasPublicApi).length;
  const e202Count = analysis.violations.filter(v => v.code === 'E202').length;
  const apiScore = Math.max(0, (slicesWithApi / analysis.slices.length) * 100 - e202Count * 5);

  // Slice Isolation: Penalize E201 violations
  const e201Count = analysis.violations.filter(v => v.code === 'E201').length;
  const isolationScore = Math.max(0, 100 - e201Count * 10);

  // Naming Consistency: Check dominant pattern match
  const namingScore = analysis.patterns.namingConsistency * 100;

  // Segment Structure: Check standard segment usage
  const segmentScore = analysis.patterns.segmentConsistency * 100;

  return Math.round(
    layerScore * weights.layerHierarchy +
    apiScore * weights.publicApi +
    isolationScore * weights.sliceIsolation +
    namingScore * weights.namingConsistency +
    segmentScore * weights.segmentStructure
  );
}

function getScoreLabel(score: number): string {
  if (score >= 90) return 'Excellent';
  if (score >= 80) return 'Good';
  if (score >= 70) return 'Fair';
  if (score >= 60) return 'Needs Work';
  return 'Critical';
}

Step 5: Generate Report

Action: Display formatted analysis report

Report template:

═══════════════════════════════════════════════════════════════
                    FSD ARCHITECTURE REPORT
═══════════════════════════════════════════════════════════════

📊 Health Score: {score}/100 ({label})

📁 Layer Summary:
┌─────────────┬─────────┬───────────┬─────────┐
│ Layer       │ Slices  │ Files     │ Status  │
├─────────────┼─────────┼───────────┼─────────┤
│ {layerName} │ {count} │ {files}   │ {status}│
└─────────────┴─────────┴───────────┴─────────┘

⚠️ Issues Found: {issueCount}

  {for each violation}
  {index}. [{code}] {source} → {target} ({description})
     Location: {file}:{line}

📈 Dependency Graph:
   {visual graph representation}

💡 Recommendations:
   {numbered list of suggestions}

═══════════════════════════════════════════════════════════════

Status icons:

✓  = No issues
⚠ N = N issues found
✗  = Critical issues

Recommendation generation:

function generateRecommendations(violations: Violation[]): string[] {
  const recommendations: string[] = [];

  // Group by violation type
  const e201s = violations.filter(v => v.code === 'E201');
  const e202s = violations.filter(v => v.code === 'E202');
  const e203s = violations.filter(v => v.code === 'E203');
  const e204s = violations.filter(v => v.code === 'E204');

  if (e201s.length > 0) {
    recommendations.push(
      `Extract shared logic between ${e201s[0].source} and ${e201s[0].target} to entities or shared layer`
    );
  }

  if (e202s.length > 0) {
    recommendations.push(
      `Export ${e202s[0].target} through its slice public API (index.ts)`
    );
  }

  if (e203s.length > 0) {
    recommendations.push(
      `Refactor ${e203s[0].source} to not depend on higher layer ${e203s[0].target}`
    );
  }

  if (e204s.length > 0) {
    recommendations.push(
      `Add index.ts to slices: ${e204s.map(v => v.source).join(', ')}`
    );
  }

  return recommendations;
}

Example output:

═══════════════════════════════════════════════════════════════
                    FSD ARCHITECTURE REPORT
═══════════════════════════════════════════════════════════════

📊 Health Score: 85/100 (Good)

📁 Layer Summary:
┌─────────────┬─────────┬───────────┬─────────┐
│ Layer       │ Slices  │ Files     │ Status  │
├─────────────┼─────────┼───────────┼─────────┤
│ app         │ -       │ 12        │ ✓       │
│ pages       │ 5       │ 45        │ ✓       │
│ widgets     │ 3       │ 28        │ ✓       │
│ features    │ 8       │ 156       │ ⚠ 2     │
│ entities    │ 4       │ 67        │ ✓       │
│ shared      │ -       │ 89        │ ✓       │
└─────────────┴─────────┴───────────┴─────────┘

⚠️ Issues Found: 2

  1. [E201] features/auth → features/user (forbidden cross-slice)
     Location: src/features/auth/model/session.ts:15

  2. [E202] features/cart → entities/product/internal
     Location: src/features/cart/api/addToCart.ts:8
     Bypasses public API

📈 Dependency Graph:
   pages ──→ widgets ──→ features ──→ entities ──→ shared
      │         │           │            │
      └─────────┴───────────┴────────────┘

💡 Recommendations:
   1. Move shared auth/user logic to entities/session
   2. Use entities/product public API (index.ts)

═══════════════════════════════════════════════════════════════

Step 6: Update Cache

Action: Save analysis results for incremental updates

1. Build cache object with:
   → Timestamp of analysis
   → File mtimes for all scanned files
   → Violation list
   → Layer/slice structure

2. Write to .fsd-architect.cache.json

3. Cache is automatically invalidated when:
   → Files are modified (mtime changes)
   → Config is changed
   → --force flag is used

Cache structure:

{
  "version": "1.0.0",
  "timestamp": "2025-01-15T10:30:00Z",
  "configHash": "abc123...",
  "files": {
    "src/features/auth/model/session.ts": {
      "mtime": 1705312200000,
      "imports": ["@entities/user", "@shared/api"]
    }
  },
  "violations": [...],
  "score": 85
}

Flags

Flag	Description	Example
--json	JSON 형식으로 출력	/fsdarch:analyze --json
--layer <name>	특정 레이어만 분석	/fsdarch:analyze --layer features
--slice <name>	특정 슬라이스만 분석	/fsdarch:analyze --slice auth
--force	캐시 무시하고 전체 스캔	/fsdarch:analyze --force
--verbose	상세 정보 출력	/fsdarch:analyze --verbose

Output Formats

Default (Terminal)

위의 예시와 같은 형식화된 텍스트 출력.

JSON (--json)

{
  "score": 85,
  "layers": {
    "app": { "files": 12, "issues": 0 },
    "pages": { "slices": 5, "files": 45, "issues": 0 },
    "features": { "slices": 8, "files": 156, "issues": 2 }
  },
  "issues": [
    {
      "code": "E201",
      "type": "forbidden-cross-slice",
      "source": "features/auth",
      "target": "features/user",
      "location": "src/features/auth/model/session.ts:15"
    }
  ],
  "recommendations": [
    "Move shared auth/user logic to entities/session"
  ]
}

Caching

Use skill: cache-manager

1. 분석 결과는 .fsd-architect.cache.json에 캐시
2. 파일 mtime 기반 증분 분석
3. --force 플래그로 캐시 무시 가능

Error Handling

E104: Config Not Found

[E104] Configuration not found

Run /fsdarch:init to initialize FSD Architect configuration.

E105: Invalid Layer Structure

[E105] Invalid layer structure

Layer 'features' contains non-slice directories:
  - src/features/utils/  (should be in shared/lib)
  - src/features/types/  (should be in shared/types)

Examples

Example 1: Full Analysis

/fsdarch:analyze

> Loading configuration...
> Scanning 6 layers, 20 slices...
> Analyzing dependencies...
> Health Score: 85/100

Example 2: Single Layer

/fsdarch:analyze --layer features

> Analyzing features layer...
> Found 8 slices: auth, cart, checkout, favorites, orders, profile, search, wishlist
> Issues: 2 cross-slice imports

Example 3: JSON Output

/fsdarch:analyze --json > fsd-report.json

> Report saved to fsd-report.json