Flags hallucination risks and absolute claims in AI-generated text.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.