Supervisor Multi-Agent System (LangGraph)

A supervisor multi-agent system for autonomous research, reasoning and code execution, built on LangGraph. A delegating supervisor routes each turn to a specialist agent, either web research or sandboxed code execution, and a finalizer turns the result into an exact-match answer.

The near-term target is GAIA Level 1, with the architecture designed toward Level 3. The repository ships a command-line runner (agent doctor, agent run, agent submit) and a Gradio UI, with CI, CodeQL and security scanning on every push.

There is no hosted deployment; it runs locally. Documentation covers the architecture, configuration and decision records.