Detects suspicious and unsafe patterns in code or data using AI.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.