Our review
This skill automates the generation of various documentation types (API, code, README, architecture) by orchestrating parallel agent tasks for efficient and comprehensive documentation.
Strengths
- Parallel documentation generation
- Covers multiple documentation styles (API, code, README, architecture)
- Ensures consistency across documents
- Reduces manual effort significantly
Limitations
- Requires a clear project structure
- May need manual review for accuracy
- Complex to set up for small projects
Use this skill when you need to generate or update comprehensive documentation for a large project with multiple components.
Do not use it for very small projects or when documentation requirements are minimal and direct writing is faster.
Security analysis
SafeThe skill describes documentation orchestration patterns and includes example shell commands and Python code for parallel task management. No destructive, obfuscated, or exfiltrating actions are present. The commands use a specific CLI tool (cc-mirror) for task management, which does not pose an inherent security risk.
No concerns found
Examples
Document all REST API endpointsAdd documentation to the utils moduleCreate a README for this projectDocumentation Orchestration Patterns
┌─────────────────────────────────────────────────────────────┐
│ │
│ Good documentation is parallel-friendly. │
│ Multiple sections, generated simultaneously. │
│ │
└─────────────────────────────────────────────────────────────┘
Load when: API documentation, code documentation, README generation, architecture docs, user guides Common patterns: Endpoint Discovery, Batch JSDoc Generation, Comprehensive README
Table of Contents
API Documentation
Pattern: Endpoint Discovery and Documentation
User Request: "Document all REST API endpoints"
Phase 1: EXPLORE
└─ Explore agent: Find all route definitions
Phase 2: FAN-OUT (Parallel documentation by domain)
├─ Agent A: Document auth endpoints
├─ Agent B: Document user endpoints
├─ Agent C: Document product endpoints
└─ Agent D: Document order endpoints
Phase 3: REDUCE
└─ General-purpose agent: Compile into unified OpenAPI/Swagger spec
Pattern: Interactive Documentation
Phase 1: PIPELINE (Foundation)
├─ Explore agent: Extract all endpoints with types
└─ General-purpose agent: Generate OpenAPI spec
Phase 2: FAN-OUT (Enhancement)
├─ Agent A: Add example requests
├─ Agent B: Add example responses
└─ Agent C: Add authentication examples
Phase 3: PIPELINE
└─ General-purpose agent: Setup Swagger UI / Redoc
Code Documentation
Pattern: Batch JSDoc/Docstring Generation
User Request: "Add documentation to the utils module"
Phase 1: EXPLORE
└─ Explore agent: Find all undocumented functions
Phase 2: MAP (Parallel documentation)
├─ Agent A: Document file1.ts functions
├─ Agent B: Document file2.ts functions
└─ Agent C: Document file3.ts functions
Phase 3: PIPELINE
└─ General-purpose agent: Verify consistency, generate type docs
Pattern: Module Overview Generation
Phase 1: EXPLORE
└─ Explore agent: Map module structure, exports, dependencies
Phase 2: PIPELINE
├─ General-purpose agent: Write module overview
├─ General-purpose agent: Document public API
└─ General-purpose agent: Add usage examples
README Generation
Pattern: Comprehensive README
User Request: "Create a README for this project"
Phase 1: FAN-OUT (Parallel information gathering)
├─ Explore agent: Project structure and technologies
├─ Explore agent: Build and run scripts (package.json, Makefile)
├─ Explore agent: Environment variables and config
├─ Explore agent: Test setup and commands
└─ Explore agent: Existing docs and comments
Phase 2: REDUCE
└─ General-purpose agent: Synthesize into structured README
Sections:
- Overview and purpose
- Quick start
- Installation
- Configuration
- Usage examples
- Development setup
- Testing
- Contributing
Architecture Documentation
Pattern: C4 Model Documentation
User Request: "Document the system architecture"
Phase 1: FAN-OUT (Parallel level documentation)
├─ Agent A: Context diagram (system + external actors)
├─ Agent B: Container diagram (applications, data stores)
├─ Agent C: Component diagram (internal components)
└─ Agent D: Code diagram (critical classes/modules)
Phase 2: REDUCE
└─ General-purpose agent: Compile into architecture doc with diagrams
Pattern: Decision Record Generation
Phase 1: EXPLORE
└─ Explore agent: Find architectural patterns in code
Phase 2: FAN-OUT
├─ Agent A: Document decision 1 (why this database?)
├─ Agent B: Document decision 2 (why this framework?)
└─ Agent C: Document decision 3 (why this structure?)
Each ADR includes:
- Context
- Decision
- Consequences
- Alternatives considered
User Guides
Pattern: Feature-Based Guides
User Request: "Write user documentation for the dashboard"
Phase 1: EXPLORE
└─ Explore agent: Map dashboard features and capabilities
Phase 2: FAN-OUT (Parallel feature guides)
├─ Agent A: Guide for feature 1 (with screenshots)
├─ Agent B: Guide for feature 2
├─ Agent C: Guide for feature 3
└─ Agent D: Troubleshooting guide
Phase 3: REDUCE
└─ General-purpose agent: Compile into user manual with TOC
Task Management for Documentation
Structure documentation work with parallel generation:
# Create documentation tasks
npx cc-mirror tasks create --subject "Audit existing docs" --description "Review current documentation state..."
npx cc-mirror tasks create --subject "Document API endpoints" --description "REST API documentation..."
npx cc-mirror tasks create --subject "Document components" --description "React component docs..."
npx cc-mirror tasks create --subject "Document utilities" --description "Helper function docs..."
npx cc-mirror tasks create --subject "Review consistency" --description "Ensure consistent style..."
npx cc-mirror tasks create --subject "Verify examples" --description "Test all code examples..."
# Parallel doc generation after audit
npx cc-mirror tasks update 2 --add-blocked-by 1
npx cc-mirror tasks update 3 --add-blocked-by 1
npx cc-mirror tasks update 4 --add-blocked-by 1
npx cc-mirror tasks update 5 --add-blocked-by 2,3,4
npx cc-mirror tasks update 6 --add-blocked-by 5
# Spawn parallel documentation agents (sonnet for well-structured work)
Task(subagent_type="general-purpose", prompt="Task 2: Document API endpoints...",
model="sonnet", run_in_background=True)
Task(subagent_type="general-purpose", prompt="Task 3: Document components...",
model="sonnet", run_in_background=True)
Task(subagent_type="general-purpose", prompt="Task 4: Document utilities...",
model="sonnet", run_in_background=True)
Output Formats
| Doc Type | Format | Tool | | ------------ | ------------------- | ---------------------- | | API docs | OpenAPI/Swagger | YAML/JSON | | Code docs | JSDoc/docstrings | Inline | | READMEs | Markdown | .md files | | Architecture | Markdown + diagrams | Mermaid/PlantUML | | User guides | Markdown/HTML | Static site generators |
─── ◈ Documentation ─────────────────────
API Documentation Generator
Documentation
Automatically generates OpenAPI/Swagger API documentation.
Technical Writer
Documentation
Writes clear technical documentation following top style guides.
Documentation Maintenance
Documentation
This skill provides a structured workflow for updating project documentation including CLAUDE.md, README, and CHANGELOG. It walks through phases like inventorying existing docs, analyzing git history for needed changes, optimizing for AI readability, and ensuring cross-document consistency. Use it when synchronizing documentation with code changes or improving documentation effectiveness for AI coding agents.