After the hugging face fiasco I wanted to do something akin to mechanistic interpretability for agent swarms and multi agent collaborations. This is a shared knowledge/goal seeking platform with observability and provenance for agent and agentic groups attempting tasks