Self-governance architecture for AI agents: membrane security, epistemic hygiene, cognitive architecture, and multi-agent trust.
