The Dave Problem: Preventing Knowledge Silos in AI Projects

How to maintain architectural coherence when using agents to generate massive code volume. For example https://github.com/Open-Cap-Stack/opencap has a well