Hatch

Hatch is the official package manager for the Hatch! ecosystem. It provides powerful tools for managing MCP server packages, environments, and interacting with the Hatch registry. Its primary objective is to service Hatchling (interactive CLI-based chat application integrating local Large Language Models with MCP for tool calling), but can of course be included in other projects.

Project Update Summary

May 27, 2025: Version 0.2.1 release of the Hatch package manager ecosystem! 🎉

• Integration with Hatchling for MCP servers package management while chatting with LLMs
• Environment isolation system for managing different sets of MCP server packages
• Registry integration for discovering and installing packages
• Package template generation for creating new Hatch MCP server packages

Features

• Environment Management: Create isolated environments for different sets of MCP servers
• Package Installation: Install packages from the registry or local directories
• Dependency Resolution: Automatically resolve and manage package dependencies
• Template Generation: Create new Hatch MCP server package templates with a single command
• Package Validation: Ensure packages conform to the Hatch schema standards

Installation

# From source
git clone https://github.com/CrackingShells/Hatch.git
cd Hatch
pip install -e .

Usage

Hatch provides both a command-line interface and a Python API for integration into other tools like Hatchling.

Command Line Interface

The table below summarizes the available commands with their arguments:

Command	Description	Arguments	Example
create	Create a new package template	name - Package name --dir, -d - Target directory --description, -D - Package description	hatch create my-package --description "My awesome MCP server package"
validate	Validate a package against schema	package_dir - Path to package directory	hatch validate ./my-package
env list	List all available environments	None	hatch env list
env create	Create a new environment	name - Environment name --description, -D - Environment description	hatch env create my-env --description "Environment for biology tools"
env use	Set the current active environment	name - Environment name	hatch env use my-env
env remove	Remove an environment	name - Environment name	hatch env remove my-env
env current	Show the current environment	None	hatch env current
package add	Add a package to an environment	package_path_or_name - Path or name of package --env, -e - Environment name --version, -v - Package version	hatch package add ./my-package --env my-env hatch package add awesome-package --env my-env
package list	List packages in an environment	--env, -e - Environment name	hatch package list --env my-env
package remove	Remove a package from an environment	package_name - Name of package to remove --env, -e - Environment name	hatch package remove awesome-package --env my-env

Python API

from hatch import HatchEnvironmentManager, create_package_template

# Create a new package template
create_package_template("my-package", target_dir="./packages", description="My awesome package")

# Manage environments
env_manager = HatchEnvironmentManager()
env_manager.create_environment("my-env", description="My testing environment")
env_manager.add_package("my-package", env_name="my-env")
packages = env_manager.list_packages("my-env")

Creating Packages

Hatch makes it easy to create new MCP server packages:

hatch create my-package --description "My MCP server package"

This creates a template with the following structure:

my-package/
├── __init__.py
├── server.py
├── hatch_metadata.json
└── README.md

Edit the server.py file to define your MCP tools:

import logging
from hatchling import HatchMCP

# Initialize MCP server with metadata
hatch_mcp = HatchMCP("my-package",
                 origin_citation="Your Name, 'Original Software', Year",
                 mcp_citation="Your Name, 'MCP Implementation', Year")

@hatch_mcp.tool()
def my_tool(param1: str, param2: int) -> str:
    """Description of what your tool does.
    
    Args:
        param1 (str): First parameter description.
        param2 (int): Second parameter description.
        
    Returns:
        str: Description of the return value.
    """
    hatch_mcp.logger.info(f"Tool called with {param1} and {param2}")
    return f"Processed {param1} with value {param2}"

if __name__ == "__main__":
    hatch_mcp.run()

Dependencies

Hatch depends on the following Python packages:

• jsonschema 4.0.0 or higher
• requests 2.25.0 or higher
• packaging 20.0 or higher
• Hatch-Validator

Development

Project Structure

• hatch/: Core package source code
  • cli_hatch.py: CLI implementation
  • environment_manager.py: Environment management functionality
  • package_loader.py: Package loading and installation
  • registry_retriever.py: Registry interaction
  • registry_explorer.py: Package search and discovery
  • template_generator.py: Package template generation

Related Repositories

Hatch is part of the larger Hatch! ecosystem which includes:

• Hatchling: Interactive CLI-based chat application integrating local Large Language Models with MCP for tool calling
• Hatch-Schemas: JSON schemas for package metadata and validation
• Hatch-Validator: Validation tools for Hatch packages
• Hatch-Registry: Package registry for Hatch packages

License

This project is licensed under the GNU Affero General Public License v3