This is something that I've been hacking on lately to provide some structure around how we can devise multi-agent LLM architectures. This is a work-in-progress and I'm curious to get y'alls thoughts in this wide open design space :)