BeehiveInnovations/pal-mcp-server ? reverse-engineered prompt

I need a Python server that acts as a central hub for all the major AI models. The idea is to connect my favorite coding assistant, like the Claude CLI or Cursor, to multiple AIs at once so I can orchestrate them like my own personal AI dev team.

For instance, I should be able to ask Claude to get a code review from both Gemini Pro and GPT-5, have them discuss the best approach, and then hand off the work to Gemini to implement the changes since it has a bigger context window. The server needs to maintain conversation history across all these different models so they remember what the others said and can build on each other's work.

It should also let me launch a separate, isolated AI agent for a specific task, like a deep security audit, without messing up my main chat context. Make sure it can connect to models from OpenAI, Google, Anthropic, Grok, and also local models through Ollama.