How do I prevent an autonomous agent from performing an action it is not authorized to perform?

Do not let a model call consequential systems directly. HIOP establishes identity, evaluates authority before the action, enforces policy, and creates verifiable evidence of what occurred. A model asserting permission is not authority. If authority cannot be proven, the action does not proceed.

Deployment scope is determined during evaluation. Hood does not prove that AI will always be safe.

If this problem matches your environment, START with the systems, effects and authority boundary you need to control.