Network Time Security authenticates NTP, but an AI agent audit trail also needs synchronization health, explicit time semantics, and evidence binding.
OpenAI's Australia incident shows how AI agent governance must link task authority, runtime evidence, disclosure, remediation, and closure.
OpenAI's DNS egress incident shows why AI agent sandbox containment must cover resolvers, transitive services, telemetry, and automatic shutdown.
A capability matrix for agent security hooks that separates visibility, evidence quality, activation, failure behavior, and enforcement.