Performs a paid preflight security check for AI requests, validating inputs, policies, and compliance before execution.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.