I've kinda been working on this tooling for linking up agents and having them work together. I am hoping to make a more "hackable" or hookable system than autogen. There are probably a few projects out there like this by now. But I've made some good progress on the "hub" allowing people to share their models and styles of LLM interactions. Curious what everyone thinks and if there are any suggestions. Thank you!