AIMCP

AIMCP is an MCP (Model Context Protocol) server that enables teams to distribute MCP tool specifications from GitLab repositories to IDEs like Cursor and VS Code. The server discovers tools.json files in repositories, dynamically generates MCP tools from their specifications, and serves referenced files as resources through a secure URI scheme.

Features

• GitLab Integration: Automatically discovers and fetches tools.json files from configured repositories
• Dynamic Tool Generation: Creates MCP tools from specifications at runtime
• Secure Resource Access: Serves repository files via aimcp:// URI scheme with access control
• Dual Caching: Caches both tool specifications and file content for optimal performance
• Conflict Resolution: Handles duplicate tool names across repositories with configurable strategies
• Async Architecture: Built on fastmcp for high-performance async operations

Installation

pip install aimcp

Quick Start

1. Create a configuration file config.yaml:

gitlab:
  base_url: "https://gitlab.example.com"
  token: "your-gitlab-token"
  repositories:
    - path: "team/tools-repo"
      branch: "main"

cache:
  type: "memory"
  ttl: 3600

conflict_resolution: "prefix"

2. Run the server:

aimcp

3. Connect your IDE to the MCP server at the configured endpoint.

Tool Specifications

Each repository must contain a tools.json file following the MCP specification. Example:

{
  "tools": [
    {
      "name": "analyze_code",
      "description": "Analyze code quality and suggest improvements",
      "resourceRefs": ["analyzer_script", "analyzer_rules"]
      }
    }
  ],
  "resources": [
    {
      "name": "analyzer_script",
      "uri": "scripts/analyze.py",
    },
    {
      "name": "analyzer_rules",
      "uri": "configs/rules.yaml"
    }
  ]
}

Configuration

Conflict Resolution Strategies

• prefix: Add repository prefix to tool names (repo1_toolname, repo2_toolname)
• priority: First repository in configuration order wins
• error: Fail startup with detailed conflict report
• merge: Combine tool descriptions and resource lists

Cache Options

• memory: In-memory caching (default)
• redis: Redis-backed caching for distributed setups (not implemented)

Development

Prerequisites

• Python 3.13+
• uv (recommended package manager)

Setup

# Clone the repository
git clone https://github.com/yourusername/aimcp.git
cd aimcp

# Install dependencies
uv sync --dev

# Run tests
uv run pytest

# Run linting
uv run ruff check .

# Run type checking
uv run mypy .

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contributing

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

Vibe coding product

Generated with Claude