Integrations

  • Creates isolated Git branches to test different debugging hypotheses in parallel, validating fixes without affecting the main codebase

  • Integrates with GitHub repositories to enable debugging support, with example solutions for real GitHub issues

  • Runs as a Node.js-based MCP server to provide debugging services to AI coding agents

Deebo: Your AI Agent's Debugging Partner

Deebo is an autonomous debugging system that AI coding agents (Claude, Cline, Cursor, etc.) can delegate tricky bugs to using the Model Context Protocol (MCP). It runs structured investigations in parallel Git branches to test hypotheses, validate fixes, and helps you move faster. If your main coding agent is like a single-threaded process, Deebo introduces multi-threadedness to your development workflow.

feedback, questions/support? dm me on x @sriramenn or open an issue here

If you think your team can benefit from Deebo, we’d love to hear from you. We’re partnering with teams who use AI agents to write production code and want to maximize their productivity. Reach out for a live walkthrough, custom setup support, or to explore early access to enterprise features.

40-second sped-up video of Deebo in action on a real codebase

Deebo scales to production codebases, too. Here's an example of Deebo solving the test53 linearizer failure $100 tinygrad bug bounty by spawning 17 scenario agents and coming up with 2 valid fixes. Check out progress.md for just the solution.

🚀 Quick Install

npx deebo-setup

That's it! Follow the prompts to configure your API key and you're ready to go. Works with:

  • VS Code Agent Mode
  • Cline
  • Claude Desktop

show us you're alive!!

npx deebo-setup ping

Cursor users: https://cursor.directory/mcp/deebo

If you're not using Cline or Claude Desktop, follow these steps:

  1. Clone the repo:
    git clone https://github.com/snagasuri/deebo-prototype.git cd deebo-prototype
  2. Install dependencies:
    npm install npm run build
  3. Install required MCP tools:
    # Install uv/uvx curl -LsSf https://astral.sh/uv/install.sh | sh # Install git-mcp uvx mcp-server-git --help # Install desktop-commander npx @wonderwhy-er/desktop-commander@latest setup
  4. Configure your MCP client to use Deebo

MCP Configuration

{ "mcpServers": { "deebo": { "autoApprove": [], "disabled": false, "timeout": 30, "command": "node", "args": [ "--experimental-specifier-resolution=node", "--experimental-modules", "--max-old-space-size=4096", "/absolute/path/to/deebo/build/index.js" ], "env": { "NODE_ENV": "development", "USE_MEMORY_BANK": "true", "MOTHER_HOST": "openrouter", "MOTHER_MODEL": "anthropic/claude-3.5-sonnet", "SCENARIO_HOST": "openrouter", "SCENARIO_MODEL": "anthropic/claude-3.5-sonnet", "OPENROUTER_API_KEY": "sk-or-v1-..." }, "transportType": "stdio" } } }

Prerequisites

  • Git: For version control
  • Node.js: v18+ (includes npm)
  • Python: 3.10+ (for git-mcp)

Configuration Files

  • VS Code: ~/Library/Application Support/Code/User/settings.json
  • Cline: ~/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json
  • Claude Desktop: ~/Library/Application Support/Claude/claude_desktop_config.json

LLM Support

Deebo supports OpenRouter, Anthropic, OpenAI SDK, and Gemini models. Configure via environment variables:

  • MOTHER_HOST: LLM provider for mother agent
  • SCENARIO_HOST: LLM provider for scenario agents
  • [PROVIDER]_API_KEY: API key for chosen provider
  • Any other OpenAI-compatible API endpoint
    • OPENAI_API_KEY to your API key (e.g., 'ollama' for Ollama)
    • OPENAI_BASE_URL to your API endpoint (e.g., 'http://localhost:11434/v1' for Ollama)

See src/util/agent-utils.ts for supported models and configuration details.

📜 License

This project is licensed under the Apache License, Version 2.0 - see the LICENSE file for details.

Copyright 2025 Sriram Nagasuri

You must be authenticated.

A
security – no known vulnerabilities
A
license - permissive license
A
quality - confirmed to work

local-only server

The server can only run on the client's local machine because it depends on local resources.

Deebo is an autonomous debugging system that AI coding agents (Claude, Cline, Cursor, etc.) can delegate tricky bugs to using the Model Context Protocol (MCP). It runs structured investigations in parallel Git branches to test hypotheses, validate fixes, and helps you move faster.

  1. 🚀 Quick Install
    1. MCP Configuration
    2. Prerequisites
    3. Configuration Files
    4. LLM Support
  2. 📜 License

    Related MCP Servers

    • -
      security
      A
      license
      -
      quality
      This is a server that lets your LLMs (like Claude) talk directly to your BigQuery data! Think of it as a friendly translator that sits between your AI assistant and your database, making sure they can chat securely and efficiently.
      Last updated -
      1
      215
      37
      JavaScript
      MIT License
    • A
      security
      A
      license
      A
      quality
      This server facilitates interaction with Keboola's Storage API, enabling users to browse and manage project buckets, tables, and components efficiently through Claude Desktop.
      Last updated -
      7
      10
      Python
      MIT License
      • Apple
    • -
      security
      A
      license
      -
      quality
      A server that allows Claude to control audio playback on your computer, supporting MP3, WAV, and OGG files with features like play, list, and stop commands.
      Last updated -
      1
      Python
      MIT License
      • Apple
      • Linux
    • A
      security
      A
      license
      A
      quality
      A server that bridges Anthropic's Claude AI with Geekbot's standup management tools, allowing users to access and utilize Geekbot data within Claude conversations.
      Last updated -
      2
      7
      Python
      MIT License

    View all related MCP servers

    ID: z7fszrff4o