Scores policy violations by severity for AI moderation and compliance workflows.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.