Validates AI agent action plans for order, prerequisites, irreversible steps, and safety risks.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.