ConsultingAgents MCP Server
@MatthewPDingle
An MCP server that interfaces with OpenAI, Google, and Anthropic's APIs to give Claude Code "coworkers" to help it on difficult problems.
Overview
What is ConsultingAgents MCP Server?
A Model Context Protocol (MCP) server that allows Claude Code to consult with additional AI agents for code and problem analysis. It provides access to four expert consultants—Darren (OpenAI), Sonny (Anthropic), Sergey (OpenAI with web search), and Gemma (Google Gemini with repository analysis)—enabling multi-model perspective on coding problems.
How to use ConsultingAgents MCP Server?
Clone the repository, create a Python 3.8+ virtual environment, install dependencies from requirements.txt, set up OPENAI_API_KEY, ANTHROPIC_API_KEY, and GOOGLE_API_KEY in a .env file, then start the server with ./start_mcp_server.sh. Register with Claude Code using claude mcp add ConsultingAgents /absolute/path/to/start_mcp_server.sh and launch Claude Code with claude --mcp-debug to access the four consulting tools.
Key features of ConsultingAgents MCP Server
- Four AI agent experts: Darren, Sonny, Sergey, Gemma.
- Supports stdio and HTTP/SSE transport protocols.
- Agents use latest models: o3-mini, Claude 3.7 Sonnet, GPT-4o, Gemini 2.5 Pro.
- Gemma analyzes entire repositories via a 1M token context window.
- Seamless integration with Claude Code via MCP.
Use cases of ConsultingAgents MCP Server
- Multi-model code review and debugging.
- Web search for documentation and examples (Sergey).
- In-depth repository analysis for feature planning (Gemma).
- Getting a second opinion from a different AI model.
- Generating comprehensive development plans from GitHub repositories.
FAQ from ConsultingAgents MCP Server
What agents are available and what models do they use?
Darren uses OpenAI o3-mini, Sonny uses Anthropic Claude 3.7 Sonnet, Sergey uses GPT-4o with web search, and Gemma uses Google Gemini 2.5 Pro with a 1M token context window.
What prerequisites are needed?
Python 3.8+, OpenAI API key, Anthropic API key, Google API key, and Claude Code CLI for integration.
How do I integrate the server with Claude Code?
Register the MCP server using claude mcp add ConsultingAgents /absolute/path/to/start_mcp_server.sh, then start Claude Code with claude --mcp-debug to use the tools.
What transport options are available?
The server supports stdio (default for direct Claude Code integration) and HTTP/SSE transport. You can set the MCP_TRANSPORT environment variable to "http" or "sse".
How do I update to a new version?
Pull the latest code, restart the server, remove the old MCP entry with claude mcp remove ConsultingAgents, and re-add with the absolute path to start_mcp_server.sh.