Resources
Notes on governing AI agents.
What we have learned building Sentinel NHI: the problems agents create for identity and security teams, and how to solve them.
Strategy
Why AI agents need their own identity layer
Single sign-on, workload identity and secret managers assume a predictable subject. Agents are not one.
FrameworkThe seven layers of agent governance
Seven questions every agent program has to answer, and what good looks like for each.
DetectionCatching attacks in the pattern, not the call
Provenance chains, honey tokens and persistent denials: how to see an attack made of allowed steps.