Scans text for leaked secrets and credential patterns, returning matched findings with severity and remediation hints.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.