Thanks for your interest. You can read "Herding Agents: A Five-Layer Architecture for Governing AI at Scale" at the link below.
What Illinois SB 315 covers, and the enterprise agent gap it leaves untouched.
Most AI security vendors can't answer the five questions that matter for runtime agent control – and the gap leaves your non-human identities exposed.
Onyx now runs as a guardrail on the TrueFoundry AI Gateway, so one security policy covers every prompt and response across every model, with no application code changes.
