MCP Roo - Proofessor AI Agent MCP Server

An MCP (Model Context Protocol) server that allows other AI agents to communicate with Proofessor, an expert AI agent specialized in code explanation, technical concepts, debugging, and architecture analysis.

Overview

This MCP server provides a standardized interface for AI agents to ask questions and get expert explanations from Proofessor. It's designed to work with Roo Cloud and can be easily integrated into any MCP-compatible environment.

Features

The server provides five specialized tools for interacting with Proofessor:

1. ask_proofessor

Ask Proofessor any question about code, architecture, debugging, or technical concepts.

Parameters:

• question (required): Your question for Proofessor
• context (optional): Additional context like code snippets or project details

Example:

{
  "question": "What are the benefits of using TypeScript over JavaScript?",
  "context": "Working on a large-scale web application with multiple developers"
}

2. explain_code

Request detailed explanations of how specific code works.

Parameters:

• code (required): The code snippet to explain
• language (optional): Programming language (e.g., 'javascript', 'python')
• focus (optional): Specific aspect to focus on (e.g., 'performance', 'security')

Example:

{
  "code": "const debounce = (fn, delay) => { let timer; return (...args) => { clearTimeout(timer); timer = setTimeout(() => fn(...args), delay); }; }",
  "language": "javascript",
  "focus": "performance"
}

3. analyze_architecture

Get architectural analysis and recommendations.

Parameters:

• description (required): Description of the architecture or system design
• concerns (optional): Specific concerns (e.g., 'scalability', 'security')

Example:

{
  "description": "Microservices architecture with event-driven communication using message queues",
  "concerns": "scalability and fault tolerance"
}

4. debug_help

Get assistance with debugging issues and understanding errors.

Parameters:

• error_message (required): The error message or problem description
• code_context (optional): Relevant code related to the error
• attempted_solutions (optional): Solutions already tried

Example:

{
  "error_message": "TypeError: Cannot read property 'map' of undefined",
  "code_context": "const items = data.items.map(item => item.name);",
  "attempted_solutions": "Checked if data exists, but still getting the error"
}

5. best_practices

Ask about best practices for specific technologies or patterns.

Parameters:

• topic (required): The technology or pattern to get best practices for
• context (optional): Specific use case or context

Example:

{
  "topic": "React hooks",
  "context": "Building a complex form with multiple validation steps"
}

Installation

Using npx (Recommended)

The easiest way to use this server is with npx:

npx @mtuckerb/mcp-roo

Global Installation

npm install -g @mtuckerb/mcp-roo

Local Installation

npm install @mtuckerb/mcp-roo

Configuration

For Roo Cloud

Add to your MCP settings file (~/.roo/mcp_settings.json or similar):

{
  "mcpServers": {
    "proofessor": {
      "command": "npx",
      "args": ["@mtuckerb/mcp-roo"],
      "disabled": false,
      "alwaysAllow": [],
      "disabledTools": []
    }
  }
}

For Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or the equivalent on your OS:

{
  "mcpServers": {
    "proofessor": {
      "command": "npx",
      "args": ["@mtuckerb/mcp-roo"]
    }
  }
}

Using Local Build

If you've cloned and built the repository locally:

{
  "mcpServers": {
    "proofessor": {
      "command": "node",
      "args": ["/path/to/mcp_roo/build/index.js"]
    }
  }
}

Development

Prerequisites

• Node.js 18 or higher
• npm or yarn

Setup

1. Clone the repository:

git clone https://github.com/mtuckerb/mcp_roo.git
cd mcp_roo

2. Install dependencies:

npm install

3. Build the project:

npm run build

Project Structure

mcp_roo/
├── src/
│   └── index.ts          # Main server implementation
├── build/                # Compiled JavaScript output
├── package.json          # Project configuration
├── tsconfig.json         # TypeScript configuration
└── README.md            # This file

Scripts

• npm run build - Compile TypeScript to JavaScript
• npm run watch - Watch mode for development
• npm run prepare - Automatically runs on install

Usage Examples

From Another AI Agent

When configured in an MCP-compatible environment, other agents can use these tools:

// Ask a general question
await use_mcp_tool("proofessor", "ask_proofessor", {
  question: "What is the difference between async/await and Promises?",
  context: "I'm refactoring callback-based code"
});

// Get code explanation
await use_mcp_tool("proofessor", "explain_code", {
  code: "function* fibonacci() { let [a, b] = [0, 1]; while (true) { yield a; [a, b] = [b, a + b]; } }",
  language: "javascript"
});

// Analyze architecture
await use_mcp_tool("proofessor", "analyze_architecture", {
  description: "REST API with JWT authentication and PostgreSQL database",
  concerns: "security and performance"
});

Testing Locally

You can test the server using the MCP inspector or by integrating it into your development environment.

How It Works

This MCP server acts as a bridge between other AI agents and Proofessor's expertise. When a tool is called:

1. The agent sends a request through the MCP protocol
2. The server receives and validates the request
3. The request is formatted for Proofessor's analysis
4. A response is generated and returned through MCP

Note: The current implementation provides simulated responses. In a production environment, this would connect to Proofessor's actual AI system for real-time expert analysis.

Contributing

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

License

MIT

Author

mtuckerb

Links

• GitHub Repository
• Issue Tracker
• Model Context Protocol Documentation

Changelog

1.0.0 (2025-11-19)

• Initial release
• Five specialized tools for interacting with Proofessor
• Support for npx usage
• Comprehensive documentation