メニュー
Generate comprehensive architecture documentation automatically from existing codebase analysis. This skill should be used when working with brownfield projects or updating outdated documentation.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/adolfoaranaes12/BMAD-ENHANCED --skill document-projectコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
Comprehensive brownfield architecture analysis for existing codebases. Discovers structure, identifies patterns, assesses quality, calculates production readiness, and provides actionable recommendations. Use when analyzing existing codebases to understand architecture, assess quality, or prepare for modernization.
Systematically consume QA quality gate outputs and apply prioritized fixes for high-severity issues, NFR failures, and test coverage gaps. Use when quality review identifies issues requiring remediation.
Peer review architecture for quality, risks, and optimization opportunities. Analyzes scalability bottlenecks, security vulnerabilities, performance optimization, technology fit, and provides prioritized recommendations. Use when reviewing proposed architecture documents for quality assurance, risk identification, or architectural decision validation.
SOC 職業分類に基づく
SKILL.md を表示中
| name | document-project |
| description | Generate comprehensive architecture documentation automatically from existing codebase analysis. This skill should be used when working with brownfield projects or updating outdated documentation. |
| acceptance | [{"documentation_generated":"All three documentation files created (architecture.md, standards.md, patterns.md)"},{"confidence_sufficient":"Overall confidence score ≥70% across all analyzed sections"},{"review_checklist_created":"Human review checklist generated identifying low-confidence areas"},{"configuration_updated":"Project configuration updated with documentation paths and brownfield flag"}] |
| inputs | {"codebase_path":{"type":"string","required":true,"description":"Path to codebase to analyze (e.g., src/)"},"existing_docs_mode":{"type":"enum","values":["merge","replace","supplement"],"default":"merge","description":"How to handle existing documentation"},"include_tests":{"type":"boolean","default":true,"description":"Include test file analysis in documentation"},"max_files":{"type":"number","default":1000,"description":"Maximum files to analyze (safety limit)"}} |
| outputs | {"documentation_files":{"type":"object","description":"Paths to generated documentation files"},"confidence_score":{"type":"number","description":"Overall confidence score (0-100%)"},"review_checklist":{"type":"string","description":"Path to human review checklist file"},"analysis_summary":{"type":"object","description":"Summary of analysis (files analyzed, patterns found, etc.)"}} |
| telemetry | {"emit":"skill.document-project.completed","track":["project_name","codebase_path","duration_ms","files_analyzed","lines_analyzed","confidence_score","patterns_identified"]} |
Generate comprehensive architecture documentation automatically by analyzing existing codebase structure, patterns, and conventions.
Analyze an existing codebase and generate three comprehensive documentation files:
This enables BMAD Enhanced to work with brownfield projects by reverse-engineering architecture from code.
This skill should be used when:
This skill should NOT be used when:
Well-Supported Languages:
Basic Support:
Optimal Codebase Size: 10K-100K lines
Execute steps in order - each builds on previous analysis:
Purpose: Verify project is suitable for automated documentation.
Actions:
.claude/config.yaml (codebasePath, existingDocs, includeTests, maxFiles)See: references/templates.md#step-0-configuration-and-validation-output for complete formats
Halt if:
Output: Validation confirmation (project type, lines, existing docs mode, scope, ready status)
Reference: See validation-criteria.md for detailed validation rules.
Purpose: Map file organization and module structure.
Actions:
Output: Project structure analysis with type, organization, directory structure (dirs + file counts + line counts), key patterns
See: references/templates.md#step-1-codebase-analysis-output for complete format
Reference: See analysis-techniques.md for detailed analysis methods.
Purpose: Identify languages, frameworks, and dependencies.
Actions:
Read package configuration:
Extract dependencies and identify frameworks:
Detect runtime/platform:
Output: Tech stack summary (runtime, backend/frontend frameworks, key libraries, testing tools, versions, confidence score)
See: references/templates.md#step-2-technology-stack-analysis for complete format
Purpose: Document data structures and validation rules.
Actions:
Locate data model files:
prisma/schema.prismasrc/types/*.ts, src/models/*.tsprisma/migrations/, migrations/Parse data models and extract validation rules
Analyze relationships:
Detect data flow:
Output: Data models summary (models with fields/types/constraints, validation rules, relationships, confidence score)
See: references/templates.md#complete-architecture-md-template for data model format
Reference: See analysis-techniques.md for model extraction patterns.
Purpose: Document API structure and conventions.
Actions:
Locate API definitions:
src/routes/**/*.tssrc/controllers/**/*.tsExtract endpoints and analyze request/response patterns:
Identify authentication:
Detect rate limiting and other middleware
Output: API specs summary (base URL, authentication type/headers/expiry, error/success response formats, confidence score)
See: references/templates.md#complete-architecture-md-template for API specification format
Purpose: Document implicit conventions and best practices.
Actions:
Analyze code style:
.eslintrc, .prettierrc, tsconfig.jsonIdentify architectural patterns:
Extract error handling patterns:
Detect testing patterns:
Output:
Coding Standards:
Code Style:
- Indentation: 2 spaces
- Quotes: Single quotes
- Naming: camelCase (variables), PascalCase (classes), kebab-case (files)
Architectural Patterns:
1. Layered Architecture
- Routes (presentation layer)
- Services (business logic layer)
- Repositories (data access layer)
2. Dependency Injection
3. Repository Pattern
Error Handling:
- Custom error classes (AppError, ValidationError)
- Centralized error handling middleware
- Never expose stack traces to clients
Confidence: High (90%)
Reference: See pattern-detection.md for pattern identification techniques.
Purpose: Create comprehensive architecture.md from findings.
Actions:
Load architecture template structure
Populate sections with analyzed data from Steps 1-5:
Generate diagrams (optional):
Add human review notes for medium/low confidence sections
Write documentation file:
docs/architecture.mdOutput:
✓ Architecture documentation generated
✓ File: docs/architecture.md (2,450 lines)
✓ Sections: 12
✓ Overall confidence: 85%
✓ Human review items: 5
Reference: See documentation-templates.md for template structures.
Purpose: Create standards.md from discovered patterns.
Actions:
Extract standards from code analysis:
Document best practices:
Create standards document with examples and consistency scores
Output:
✓ Standards documentation generated
✓ File: docs/standards.md (850 lines)
✓ Standards extracted: 18
✓ Consistency scores: 75-100%
Purpose: Document discovered design patterns and conventions.
Actions:
Extract design patterns:
Document usage examples:
Identify anti-patterns:
Create patterns document
Output:
✓ Patterns documentation generated
✓ File: docs/patterns.md (620 lines)
✓ Patterns identified: 8
✓ Anti-patterns noted: 3
Reference: See pattern-detection.md for pattern documentation templates.
Purpose: Score accuracy and identify areas needing human review.
Actions:
Output: Validation summary (overall confidence %, high/medium priority review item counts)
See: references/templates.md#review-checklist-template and references/confidence-scoring.md for formats
Reference: See confidence-scoring.md for scoring guidelines.
Purpose: Provide user with clear summary and action items.
Actions:
docs/REVIEW_CHECKLIST.mdOutput: Summary confirmation (report generated, checklist created, config updated, ready status)
See: references/templates.md for complete summary formats
High (85-100%): Explicit in code/config, consistent patterns, no conflicts | Medium (70-84%): Inferred from patterns, some inconsistencies, needs validation | Low (<70%): Missing/unclear info, conflicts, high uncertainty, MUST review
See: references/confidence-scoring.md for detailed scoring methodology
Cannot: Understand business logic without context, document deployment infrastructure, capture tribal knowledge, understand legacy decisions, document external integrations perfectly
Requires: Readable structured codebase, standard organization, supported language/framework, 10K-100K lines (optimal)
Run periodically (every 3-6 months) | Always review low-confidence sections | Supplement with manual docs (business context, deployment) | Use as starting point, enhance with team input
Brownfield: document-project → index-docs → create-task-spec (use generated docs) | Greenfield: Write docs manually first | Brownfield: Generate from code, then refine
Detailed documentation in references/: