Detects honeypots, rug pulls, and malicious tokens with a safe/caution/risky verdict.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.