Zero Trust for Agentic Systems — trust, safety, and security for agents. @mvpzone
Six Means of Knowing
How do you know what an AI agent’s output is actually true? Any single verifier has a blind spot — it is the shape of what that instrument cannot see. Older traditions of inquiry named six independent channels for valid knowing. If you want to verify what an agent produced, you need more than one.
Verification, Not Prevention
Conventional security promises prevention. Zero Trust for agentic systems refuses this aspiration. It promises something narrower and stranger: that when failures occur, they will be bounded, attested, recoverable, and defensible. Prevention is declined because the prevention problem, in agentic systems, is not solvable in the form it is posed. Verification is.
Safety Is Bigger Than Security
In conventional thinking, safety is a subset of security: harden the system, harm doesn’t land. For agentic systems, this gets the relationship backwards. Security is a subset of safety. Authorised use — not just authorised access — determines safety.