MindRoot MCP Integration Plugin

This plugin integrates the Model Context Protocol (MCP) with MindRoot, allowing the AI to connect to and use MCP servers for enhanced capabilities.

Features

• MCP Client: Connect to MCP servers via stdio, SSE, or WebSocket
• Server Management: Add, configure, and manage MCP servers
• uvx/npx Support: Easy installation and running of MCP servers via uvx (Python) and npx (Node.js)
• Tool Integration: Use MCP tools as MindRoot commands
• Resource Access: Read resources from MCP servers
• Admin Interface: Web UI for managing MCP servers
• Server Discovery: Discover available MCP servers from online directories

Installation

1. Install the MCP Python SDK and optional tools:

pip install mcp

For uvx support: pip install uvx For npx support: Install Node.js and npm

2. Install the plugin:

cd /xfiles/update_plugins/mr_mcp
pip install -e .

3. Enable the plugin commands in the MindRoot admin interface

Usage

Admin Interface

Access the MCP management interface at /admin and look for the "MCP Servers" section.

Commands

• mcp_list_servers - List all configured MCP servers
• mcp_connect - Connect to an MCP server
• mcp_disconnect - Disconnect from an MCP server
• mcp_install_uvx_server - Install and configure a uvx-based MCP server
• mcp_install_npx_server - Install and configure an npx-based MCP server
• mcp_list_tools - List available tools from MCP servers
• mcp_call_tool - Call a tool on an MCP server
• mcp_list_resources - List available resources
• mcp_read_resource - Read a resource from an MCP server
• mcp_discover_servers - Discover available MCP servers
• mcp_install_server - Configure a new MCP server

Example Usage

// Install a uvx-based server (Python)
{ "mcp_install_uvx_server": {
    "name": "calculator",
    "package": "mcp-server-calculator",
    "description": "Calculator server"
}}

// Install an npx-based server (Node.js)
{ "mcp_install_npx_server": {
    "name": "github",
    "package": "@modelcontextprotocol/server-github",
    "description": "GitHub integration server"
}}

// Configure a filesystem server
{ "mcp_install_server": {
    "name": "filesystem",
    "description": "File system operations",
    "command": "python",
    "args": ["-m", "mcp.server.filesystem"]
}}

// Connect to the server
{ "mcp_connect": { "server_name": "filesystem" } }

// List available tools
{ "mcp_list_tools": {} }

// Call a tool
{ "mcp_call_tool": {
    "server_name": "filesystem",
    "tool_name": "read_file",
    "arguments": {"path": "/tmp/test.txt"}
}}

Popular MCP Servers

• Calculator (uvx): mcp-server-calculator - Mathematical calculations
• Filesystem (Python): File operations
• GitHub (npx): @modelcontextprotocol/server-github - GitHub API integration
• Brave Search (npx): @modelcontextprotocol/server-brave-search - Web search
• SQLite (npx): @modelcontextprotocol/server-sqlite - Database operations
• Slack (npx): @modelcontextprotocol/server-slack - Slack integration
• Google Drive (npx): @modelcontextprotocol/server-gdrive - Google Drive access
• Puppeteer (npx): @modelcontextprotocol/server-puppeteer - Web automation

See the discovery feature to find more available servers.

Architecture

• MCPManager: Core class managing server connections
• MCPServer: Pydantic model for server configuration
• Commands: MindRoot commands for MCP operations
• Admin UI: Web interface for server management
• Router: FastAPI routes for the admin interface

Configuration

Server configurations are stored in /tmp/mcp_servers.json and include:

• Server name and description
• Command and arguments
• Transport type (stdio, SSE, WebSocket)
• Connection status
• Available capabilities (tools, resources, prompts)

Development

The plugin supports multiple installation methods:

• uvx: For Python-based MCP servers (similar to pipx)
• npx: For Node.js-based MCP servers
• pip: For Python packages
• npm: For Node.js packages
• manual: For custom installations

To extend the plugin:

1. Add new commands in mod.py
2. Update the admin interface in templates/mcp_admin.jinja2
3. Add new server discovery sources
4. Implement additional transport types

Troubleshooting

• Ensure MCP SDK is installed: pip install mcp
• For uvx servers: Install uvx with pip install uvx
• For npx servers: Ensure Node.js and npm are installed
• Check server command paths and arguments
• Verify server processes are running
• Check logs for connection errors
• Ensure proper permissions for server executables