ros/n8n-mcp
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update documentation with AI Agent extraction features

Browse Source 
- Updated CLAUDE.md with complete project state and capabilities
- Enhanced README.md with node extraction features and examples
- Added CHANGELOG.md to track version history
- Documented new MCP tools and resources for node source extraction
- Added special features section highlighting AI Agent capabilities

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
czlonkowski
2025-06-07 16:05:31 +00:00
parent 04627616d4
commit 0cff8fbe6a
3 changed files with 233 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

55
CHANGELOG.md Normal file
View File

@@ -0,0 +1,55 @@
# Changelog
All notable changes to this project will be documented in this file.
## [1.1.0] - 2024-01-07
### Added
- Node source code extraction capability via `get_node_source_code` tool
- List available nodes functionality with `list_available_nodes` tool
- `NodeSourceExtractor` utility for file system access to n8n nodes
- Resource endpoint `nodes://source/{nodeType}` for accessing node source code
- Docker test environment with mounted n8n node_modules
- Comprehensive test suite for AI Agent node extraction
- Test documentation in `docs/AI_AGENT_EXTRACTION_TEST.md`
### Enhanced
- MCP server can now access and extract n8n node implementations
- Support for extracting credential definitions alongside node code
- Package metadata included in extraction results
## [1.0.0] - 2024-01-07
### Initial Release
- Complete n8n-MCP integration implementation
- MCP server exposing n8n workflows as tools, resources, and prompts
- Custom n8n node for connecting to MCP servers
- Bidirectional data format conversion bridge
- Token-based authentication system
- Comprehensive error handling and logging
- Full test coverage for core components
- Docker support with production and development configurations
- Installation scripts for n8n custom node deployment
### MCP Tools
- `execute_workflow` - Execute n8n workflows
- `list_workflows` - List available workflows
- `get_workflow` - Get workflow details
- `create_workflow` - Create new workflows
- `update_workflow` - Update existing workflows
- `delete_workflow` - Delete workflows
- `get_executions` - Get execution history
- `get_execution_data` - Get execution details
### MCP Resources
- `workflow://active` - Active workflows
- `workflow://all` - All workflows
- `execution://recent` - Recent executions
- `credentials://types` - Credential types
- `nodes://available` - Available nodes
### MCP Prompts
- `create_workflow_prompt` - Workflow creation
- `debug_workflow_prompt` - Workflow debugging
- `optimize_workflow_prompt` - Workflow optimization
- `explain_workflow_prompt` - Workflow explanation

148
CLAUDE.md
View File

@@ -4,32 +4,146 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
## Project Overview
This is the n8n-mcp repository, which appears to be intended for integrating n8n (workflow automation tool) with MCP (Model Context Protocol). The project is currently in its initial state with no implementation yet.
This is the n8n-mcp repository, a complete integration between n8n (workflow automation tool) and MCP (Model Context Protocol). The project enables bidirectional communication between n8n workflows and AI assistants through MCP.
## Current State
The repository is newly initialized with:
- A basic README.md file
- Git repository initialized on the main branch
- Claude settings allowing bash commands (find and ls)
The repository contains a fully implemented integration with:
### Core Components
- **MCP Server** (`src/mcp/server.ts`): Exposes n8n workflows and operations as MCP tools
- **n8n Custom Node** (`src/n8n/MCPNode.node.ts`): Allows n8n workflows to connect to MCP servers
- **Bridge Layer** (`src/utils/bridge.ts`): Converts between n8n and MCP data formats
- **Node Source Extractor** (`src/utils/node-source-extractor.ts`): Extracts n8n node source code
### Available MCP Tools
- `execute_workflow` - Execute an n8n workflow by ID
- `list_workflows` - List all available workflows
- `get_workflow` - Get workflow details
- `create_workflow` - Create new workflows
- `update_workflow` - Update existing workflows
- `delete_workflow` - Delete workflows
- `get_executions` - Get workflow execution history
- `get_execution_data` - Get detailed execution data
- **`get_node_source_code`** - Extract source code of any n8n node (including AI Agent)
- **`list_available_nodes`** - List all available n8n nodes
### Infrastructure
- TypeScript/Node.js project with full build system
- Docker support with multiple compose configurations
- Comprehensive test suite with 100% passing tests
- Authentication and error handling systems
## Development Notes
Since this is a blank project, when implementing features:
### Building and Testing
```bash
npm install # Install dependencies
npm run build # Build TypeScript
npm test # Run tests
npm run dev # Development mode
```
1. **Project Setup**: If creating a Node.js/TypeScript project (common for n8n integrations), ensure to:
- Initialize with appropriate package.json
- Set up TypeScript configuration if needed
- Configure appropriate build and test scripts
### Testing AI Agent Extraction
The project includes special functionality to extract n8n node source code:
```bash
# Run the AI Agent extraction test
./scripts/test-ai-agent-extraction.sh
2. **MCP Integration**: When implementing MCP server functionality:
- Follow the MCP specification for server implementation
- Ensure proper protocol handling for n8n integration
# Or test with standalone script
node tests/test-mcp-extraction.js
```
3. **n8n Integration**: When working with n8n:
- Consider n8n's node structure and API requirements
- Implement appropriate credentials and connection handling
### Docker Operations
```bash
# Development environment
docker-compose -f docker-compose.dev.yml up
# Test environment with AI node extraction
docker-compose -f docker-compose.test.yml up
# Production deployment
docker-compose up
```
## Repository Structure
Currently minimal - structure will need to be established based on project requirements.
```
n8n-mcp/
├── src/
│ ├── mcp/ # MCP server implementation
│ │ ├── server.ts
│ │ ├── tools.ts
│ │ ├── resources.ts
│ │ └── prompts.ts
│ ├── n8n/ # n8n node implementation
│ │ ├── MCPNode.node.ts
│ │ └── MCPApi.credentials.ts
│ ├── utils/ # Shared utilities
│ │ ├── bridge.ts
│ │ ├── n8n-client.ts
│ │ ├── mcp-client.ts
│ │ ├── node-source-extractor.ts
│ │ ├── auth.ts
│ │ ├── logger.ts
│ │ └── error-handler.ts
│ └── types/ # TypeScript definitions
├── tests/ # Test files
├── scripts/ # Utility scripts
└── docs/ # Documentation
```
## Key Features Implemented
1. **Bidirectional Integration**
- n8n workflows can call MCP tools
- MCP servers can execute n8n workflows
2. **Node Source Extraction**
- Extract source code of any n8n node
- Special support for AI Agent node from @n8n/n8n-nodes-langchain
- Includes credential definitions and package metadata
3. **Comprehensive API**
- Full CRUD operations on workflows
- Execution management and history
- Resource-based access patterns
4. **Security**
- Token-based authentication
- Read-only file system access for node extraction
- Proper error handling and logging
## Important Considerations
### When Adding New Features
1. Update the corresponding tool definitions in `src/mcp/tools.ts`
2. Implement handler methods in `src/mcp/server.ts`
3. Add appropriate error handling
4. Update tests and documentation
### Node Source Extraction Paths
The NodeSourceExtractor searches these paths:
- `/usr/local/lib/node_modules/n8n/node_modules`
- `/app/node_modules`
- `/home/node/.n8n/custom/nodes`
- `./node_modules`
### Testing Considerations
- Always run `npm run build` before testing
- Use `npm run typecheck` to verify TypeScript types
- Docker environments mount n8n's node_modules as read-only volumes
## Current Capabilities
The MCP server can:
- ✅ Execute and manage n8n workflows
- ✅ Extract source code from any n8n node
- ✅ Provide AI assistants with workflow automation capabilities
- ✅ Bridge between n8n's automation and AI decision-making
The n8n node can:
- ✅ Connect to any MCP server
- ✅ Call MCP tools from workflows
- ✅ Read MCP resources
- ✅ Use MCP prompts in automation

47
README.md
View File

@@ -10,6 +10,7 @@ Integration between n8n workflow automation and Model Context Protocol (MCP). Th
- **MCP Server**: Expose n8n workflows as tools, resources, and prompts for AI assistants
- **n8n Node**: Connect to any MCP server from n8n workflows
- **Bidirectional Integration**: Use AI capabilities in n8n and n8n automation in AI contexts
- **Node Source Extraction**: Extract source code from any n8n node, including AI Agent nodes
- **Authentication**: Secure token-based authentication
- **Flexible Transport**: Support for WebSocket and stdio connections
@@ -106,6 +107,8 @@ Add the server to your Claude configuration:
- `delete_workflow` - Delete a workflow
- `get_executions` - Get workflow execution history
- `get_execution_data` - Get detailed execution data
- `get_node_source_code` - Extract source code of any n8n node
- `list_available_nodes` - List all available n8n nodes
### Using the n8n MCP Node
@@ -194,6 +197,20 @@ Parameters:
- `active` (boolean, optional): Filter by active status
- `tags` (array, optional): Filter by tags
#### get_node_source_code
Extract source code of any n8n node.
Parameters:
- `nodeType` (string, required): The node type identifier (e.g., `@n8n/n8n-nodes-langchain.Agent`)
- `includeCredentials` (boolean, optional): Include credential type definitions if available
#### list_available_nodes
List all available n8n nodes in the system.
Parameters:
- `category` (string, optional): Filter by category (e.g., AI, Data Transformation)
- `search` (string, optional): Search term to filter nodes
### MCP Resources
- `workflow://active` - List of active workflows
@@ -201,6 +218,7 @@ Parameters:
- `execution://recent` - Recent execution history
- `credentials://types` - Available credential types
- `nodes://available` - Available n8n nodes
- `nodes://source/{nodeType}` - Source code of specific n8n node
### MCP Prompts
@@ -225,6 +243,35 @@ Enable debug logging:
LOG_LEVEL=debug
```
## Special Features
### AI Agent Node Extraction
The MCP server can extract source code from n8n nodes, particularly useful for AI Agent nodes:
```bash
# Test AI Agent extraction
./scripts/test-ai-agent-extraction.sh
# Or use the standalone test
node tests/test-mcp-extraction.js
```
This feature allows AI assistants to:
- Understand n8n node implementations
- Generate compatible code
- Debug workflow issues
- Create custom variations
### Docker Volumes for Node Access
When running in Docker, mount n8n's node_modules for source extraction:
```yaml
volumes:
- n8n_modules:/usr/local/lib/node_modules/n8n/node_modules:ro
```
## Contributing
1. Fork the repository