Checks whether a target shows signs of exploit risk or compromise.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.