I’m curious: how do you build such a knowledge base? It’s still not clear to me what form it should take? A simple repo with plain text files?

This matches my experience. The bottleneck isn't what the agent knows, it's what the agent can verify. A knowledge base tells it "don't do X", but the agent still has to remember to check. Giving it a tool that returns ground truth works better. The agent calls the tool, gets a concrete answer, acts on it. No memory required, no drift over time.