Skip to main content

Extend Click applications with Model Context Protocol (MCP) support

Project description

click-mcp

A Python library that extends Click applications with Model Context Protocol (MCP) support, allowing AI agents to interact with CLI tools.

Overview

click-mcp provides a simple decorator that converts Click commands into MCP tools. This enables AI agents to discover and interact with your CLI applications programmatically.

The Model Context Protocol (MCP) is an open standard for AI agents to interact with tools and applications in a structured way.

Key Features

  • Simple @click_mcp decorator syntax
  • Automatic conversion of Click commands to MCP tools
  • Support for nested command groups
  • Stdio-based MCP server for easy integration

Installation

pip install click-mcp

Basic Usage

import click
from click_mcp import click_mcp

@click_mcp
@click.group()
def cli():
    """Sample CLI application."""
    pass

@cli.command()
@click.option('--name', required=True, help='Name to greet')
def greet(name):
    """Greet someone."""
    click.echo(f"Hello, {name}!")

if __name__ == '__main__':
    cli()

When you run the MCP server, Click commands are converted into MCP tools:

  • Command greet becomes MCP tool greet
  • Nested commands use dot notation (e.g., users.create)

To invoke a command via MCP, send a request like:

{
  "type": "invoke",
  "tool": "greet",
  "parameters": {
    "name": "World"
  }
}

To start the MCP server:

$ python my_app.py mcp

Advanced Usage

Customizing the MCP Command Name

By default, click-mcp adds an mcp command to your CLI application. You can customize this name using the command_name parameter:

@click_mcp(command_name="start-mcp")
@click.group()
def cli():
    """Sample CLI application with custom MCP command name."""
    pass

With this configuration, you would start the MCP server using:

$ python my_app.py start-mcp

This can be useful when:

  • The name "mcp" conflicts with an existing command
  • You want a more descriptive command name
  • You're integrating with a specific AI agent that expects a certain command name

Working with Nested Command Groups

click-mcp supports nested command groups. When you have a complex CLI structure with subcommands, all commands are exposed as MCP tools:

@click_mcp
@click.group()
def cli():
    """Main CLI application."""
    pass

@cli.group()
def users():
    """User management commands."""
    pass

@users.command()
@click.option('--username', required=True)
def create(username):
    """Create a new user."""
    click.echo(f"Creating user: {username}")

@users.command()
@click.argument('username')
def delete(username):
    """Delete a user."""
    click.echo(f"Deleting user: {username}")

When exposed as MCP tools, the nested commands will be available with their full path using dot notation (e.g., "users.create" and "users.delete").

Handling Command Errors

When a Click command raises an exception, click-mcp captures the error and returns it as part of the MCP response. This allows AI agents to handle errors gracefully:

@cli.command()
@click.option('--filename', required=True)
def process(filename):
    """Process a file."""
    try:
        with open(filename, 'r') as f:
            content = f.read()
        click.echo(f"Processed file: {filename}")
    except FileNotFoundError:
        raise click.UsageError(f"File not found: {filename}")

If the file doesn't exist, the AI agent will receive an error message that it can present to the user or use to take corrective action.

Development

Setup

Clone the repository and install development dependencies:

git clone https://github.com/aws/click-mcp.git
cd click-mcp
pip install -e ".[dev]"

Testing

Run tests with pytest:

pytest

Run tests with coverage:

pytest --cov=click_mcp

Code Formatting

Format code with black:

black click_mcp tests

Run linting checks:

flake8 click_mcp tests

Building

Build the package:

python -m build

Documentation

Generate documentation:

cd docs
make html

Related Resources

License

MIT

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

click_mcp-0.1.0.tar.gz (15.8 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

click_mcp-0.1.0-py3-none-any.whl (18.4 kB view details)

Uploaded Python 3

File details

Details for the file click_mcp-0.1.0.tar.gz.

File metadata

  • Download URL: click_mcp-0.1.0.tar.gz
  • Upload date:
  • Size: 15.8 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.12.9

File hashes

Hashes for click_mcp-0.1.0.tar.gz
Algorithm Hash digest
SHA256 b22cca507e64ef0f9f292ff6772f85c1c5dfa92ad50b358f4919652cc24f9b55
MD5 d66bedc69be8059ce2f185be016e44b2
BLAKE2b-256 17135c85232888350872409de3f5c4dae15bb8e59f62698667807b81b878a8d5

See more details on using hashes here.

Provenance

The following attestation bundles were made for click_mcp-0.1.0.tar.gz:

Publisher: publish-to-pypi.yml on crowecawcaw/click-mcp

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file click_mcp-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: click_mcp-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 18.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.12.9

File hashes

Hashes for click_mcp-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 3001a3665bbc327d7bf47d3e1e3a64f18fdc7e3eb7406e39988c2fd77582c323
MD5 80ed90e4c44c65e2601de87b16774f4b
BLAKE2b-256 e0e8665834f98cd90d481a741c01af168cab9d9241ce749152751aad4be09de4

See more details on using hashes here.

Provenance

The following attestation bundles were made for click_mcp-0.1.0-py3-none-any.whl:

Publisher: publish-to-pypi.yml on crowecawcaw/click-mcp

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page