Evaluates humor for appropriateness by scoring jokes and flagging offensive or sensitive content with clear suitability metrics.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.