@coderai/n8n-nodes-claude-agent

This is an n8n community node that integrates Claude Code autonomous AI agent capabilities into your n8n workflows.

n8n is a fair-code licensed workflow automation platform that allows you to connect anything to everything via its open, extensible system.

Claude Code is Anthropic's AI assistant with advanced autonomous coding capabilities, able to understand codebases, edit files, run commands, and execute complex workflows.

Features

• Generate Code: Create code snippets from natural language prompts in multiple programming languages
• Execute Agent Tasks: Run autonomous coding tasks with full agent capabilities including file reading/writing, command execution, and more
• Flexible Permission Control: Choose from multiple permission modes to control what Claude can do
• Tool Selection: Fine-tune which tools Claude can access during task execution
• Environment Configuration: Customize working directory, environment variables, and system prompts

Installation

Via n8n Community Nodes (Recommended)

1. Go to Settings > Community Nodes in your n8n instance
2. Click Install a community node
3. Enter @coderai/n8n-nodes-claude-agent in the npm package name field
4. Click Install

Manual Installation

To install this node manually in your local n8n instance:

npm install @coderai/n8n-nodes-claude-agent

For Docker-based n8n installations, you can add this package to your package.json or install it via the n8n UI.

Prerequisites

• An Anthropic API key
• n8n version 1.0.0 or above

Credentials

This node requires Claude API credentials:

1. Go to Anthropic Console
2. Create a new API key
3. In n8n, create new "Claude API" credentials
4. Paste your API key

Operations

1. Generate Code

Generate code snippets from natural language descriptions.

Parameters:

• Prompt (required): Describe what code you want to generate
• Language: Select the programming language (Python, JavaScript, TypeScript, Java, C++, C#, Go, Rust, PHP, Ruby, or Other)

Example:

Prompt: "Create a function that validates email addresses using regex"
Language: Python

Output:

{
  "code": "import re\n\ndef validate_email(email):\n    pattern = r'^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}$'\n    return re.match(pattern, email) is not None",
  "language": "python",
  "prompt": "Create a function that validates email addresses using regex",
  "messages": [...]
}

2. Execute Agent Task

Execute autonomous coding tasks with full agent capabilities.

Parameters:

• Task Description (required): Describe the autonomous task
• Working Directory: Directory where Claude will operate (defaults to current directory)
• Allowed Tools: Select which tools Claude can use:
  • Read: Read files
  • Write: Write/create files
  • Edit: Edit existing files
  • Execute: Run commands
  • Grep: Search in files
  • Glob: Find files by pattern
  • List: List directory contents
• Permission Mode: Control permission behavior:
  • Default: Standard behavior with prompts
  • Accept Edits: Auto-accept file edits
  • Bypass Permissions: Bypass all permission checks (⚠️ use with caution)
  • Plan: Planning mode without execution
• Max Turns: Maximum conversation turns (0 for unlimited)

Additional Options:

• System Prompt: Custom instructions for Claude
• Max Thinking Tokens: Limit thinking process tokens
• Model Override: Choose between Sonnet (default), Opus, or Haiku
• Environment Variables: JSON object with environment variables
• Include Partial Messages: Include partial message events in output

Example:

Task Description: "Analyze the authentication module and refactor it to use JWT tokens"
Working Directory: "/path/to/project"
Allowed Tools: Read, Write, Edit, Grep
Permission Mode: Accept Edits

Output:

{
  "task": "Analyze the authentication module and refactor it to use JWT tokens",
  "summary": "Completed refactoring of authentication to JWT...",
  "messages": [...],
  "artifacts": [...],
  "toolCalls": [
    {"tool": "Read", "input": {...}},
    {"tool": "Edit", "input": {...}}
  ],
  "workingDirectory": "/path/to/project",
  "permissionMode": "acceptEdits",
  "allowedTools": ["Read", "Write", "Edit", "Grep"]
}

Usage Examples

Example 1: Generate API Client Code

Operation: Generate Code
Prompt: "Create a REST API client class with GET, POST, PUT, DELETE methods including error handling"
Language: TypeScript

Example 2: Automated Code Review

Operation: Execute Agent Task
Task Description: "Review all Python files in the src/ directory and create a report of code quality issues, potential bugs, and improvement suggestions"
Working Directory: /path/to/project
Allowed Tools: Read, Grep, List
Permission Mode: Plan

Example 3: Refactoring Task

Operation: Execute Agent Task
Task Description: "Refactor the user authentication module to implement OAuth 2.0"
Working Directory: /path/to/project/backend
Allowed Tools: Read, Write, Edit, Grep, Glob
Permission Mode: Accept Edits
Additional Options:
  - System Prompt: "Follow the project's coding standards defined in CONTRIBUTING.md"
  - Max Turns: 20

Example 4: Test Generation

Operation: Execute Agent Task
Task Description: "Generate comprehensive unit tests for all functions in the utils.js file"
Working Directory: /path/to/project
Allowed Tools: Read, Write, Grep
Permission Mode: Accept Edits

Permission Modes Explained

• Default: Claude will request permission for sensitive operations. Best for interactive use.
• Accept Edits: Automatically approves file edits but may still prompt for other operations. Good for refactoring tasks.
• Bypass Permissions: Skips all permission checks. ⚠️ Use only in controlled environments where you trust the task completely.
• Plan: Claude analyzes and plans but doesn't execute. Perfect for understanding what Claude would do before running.

Security Considerations

⚠️ Important: When using this node, be aware that:

1. Claude Code can read, write, and execute files in the specified working directory
2. Use Bypass Permissions mode only in controlled, isolated environments
3. Always review the Allowed Tools to limit Claude's capabilities
4. Consider using Plan mode first to preview actions
5. Keep your API key secure and never commit it to version control

Troubleshooting

Node not appearing in n8n

• Ensure n8n is restarted after installation
• Check that the package is properly installed: npm list n8n-nodes-claude-code

API Key errors

• Verify your API key is valid at Anthropic Console
• Ensure credentials are properly configured in n8n

Permission errors

• Check that the working directory is accessible
• Verify file permissions in the target directory
• Try using a less restrictive permission mode for testing

Task timeout or incomplete

• Increase Max Turns if tasks are stopping prematurely
• Simplify complex tasks into smaller steps
• Check if the working directory is correct

Resources

• n8n Documentation
• Claude Agent SDK Documentation
• Anthropic API Reference
• n8n Community Forum

Development

Setup

# Clone the repository
git clone https://github.com/MyCoderAI/n8n-nodes-claude-agent.git
cd n8n-nodes-claude-agent

# Install dependencies
npm install

# Build the node
npm run build

# Link for local testing
npm link

Testing Locally

# In your n8n installation directory
npm link @coderai/n8n-nodes-claude-agent

# Start n8n
n8n start

Linting

npm run lint
npm run lint:fix

Version History

0.1.0

• Initial release
• Generate Code operation
• Execute Agent Task operation with full configuration options
• Support for multiple permission modes and tool selection

License

MIT

Support

For issues, questions, or contributions:

• GitHub Issues
• n8n Community

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

Author

CoderAI - GitHub

---

Built with ❤️ for the n8n community