Resources
AI Agent Authorization vs IAM / OAuth
Identity and access management tells you which principal may reach an API. OAuth and related protocols issue tokens and scopes. Those controls are necessary. They are not sufficient once the caller is an agent that can compose many tool calls from a single session.
A valid token answers “is this caller allowed to use the payment tool?” Decision Authorization answers “may this specific $250,000 release proceed under current intent, evidence, and authority?” A mutated $260,000 request is a new action even if the token is unchanged.
OntoGuard does not replace IAM. It sits after identity has been established and before protected execution. Enforcement systems consume ALLOW, BLOCK, or ESCALATE at the endpoint they already own.
See also governance at the runtime boundary.
Continue: Proof Explorer · Inspect the AgenTrust TRACE integration · Explore the Headless Decision Authorization Runtime.
FAQ
Does OntoGuard replace IAM?
No. IAM answers who may call a system. Decision Authorization answers whether this exact action may proceed.
Does a valid token authorize the payment?
A token authorizes use of a tool. It does not bind a specific amount, target, or duration.