Checks whether an agent task has clear objective, authority, data access, output, and success criteria.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.