project-dumper

name	project-dumper
version	1.1.0
description	Extract project knowledge into structured Markdown documents. Analyzes source code to generate architecture docs, API docs, database design, and flow diagrams. Use when: "dump project", "extract project docs", "generate project knowledge", "项目文档", or when you need to extract project structure and design.
allowed-tools	["Bash","Read","Write","Glob","Grep"]

project-dumper

Generate structured Markdown documentation from project source code. This skill helps extract project knowledge (architecture, API, database, flows) without copying source code.

Usage

/project-dumper [--output DIR] [--project PATH]

Examples:
/project-dumper                                    # Use current directory
/project-dumper --output ./docs                   # Output to ./docs
/project-dumper --project /path/to/project        # Analyze specific project
/project-dumper -o ./my-docs -p ./my-project    # Full options

Parameters

Parameter	Short	Description	Default
--output	-o	Output directory for generated docs	./project-docs
--project	-p	Project directory to analyze	Current directory

Workflow

Follow these phases to analyze the project:

Phase 1: Project Structure Analysis

1. Scan the project directory to understand:

   • Project type (language, framework)
   • Directory structure
   • Key configuration files (package.json, pom.xml, requirements.txt, etc.)
   • Module/package organization

2. Identify the main entry points and key modules

Phase 2: API & Database Analysis

For each API layer found (REST controllers, routes, endpoints):

• List all endpoints with HTTP method, path, and description
• Document request/response formats

For each database layer found (entities, models, schemas):

• List all entities/tables with fields and relationships
• Document primary keys, indexes, and constraints

Phase 3: Architecture & Design Analysis

Analyze and document:

• Layered architecture (presentation, business, data, etc.)
• Design patterns used
• Technology stack and why
• Key architectural decisions

Phase 4: Core Flows & Diagrams

Document key business flows:

• User interactions
• Data processing pipelines
• External integrations

Generate Mermaid diagrams for:

• System architecture
• Data flow
• Sequence diagrams for key operations

Output Format

Generate these Markdown files in the output directory:

1. README.md - Project overview, tech stack, quick start
2. ARCHITECTURE.md - Architecture design, layers, patterns
3. API.md - API endpoints documentation
4. DATABASE.md - Database schema and models
5. FLOWS.md - Business flows with Mermaid diagrams

Guidelines

• DO NOT copy source code - Only summarize and describe
• Use Mermaid diagrams for visual representation
• Include code snippets only as examples, not full files
• Focus on design intent - Why was it built this way?
• Be concise - Use bullet points and tables
• Language: Use the same language as the project comments/docs, default to English
• Depth: For architecture docs, include specific code examples, configuration, and implementation details that enable reverse engineering

Output Directory

Default: ./project-docs/

Custom output:

# Example 1: Output to current directory's docs folder
/project-dumper -o ./docs

# Example 2: Output to home directory
/project-dumper -o ~/project-docs

# Example 3: Absolute path
/project-dumper -o /Users/name/Documents/my-project-docs

Note: The output directory will be created if it doesn't exist.

Edge Cases

• If project has no clear API layer: Document what interfaces the project exposes
• If project has no database: Skip DATABASE.md
• If project is too large: Prioritize main modules, note that full analysis requires multiple runs
• If language/framework not recognized: Ask user for clarification
• If output directory is not writable: Report error with suggested fix