Debugs AI agent execution traces for failures, context breaks, latency bottlenecks, and injection attempts.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.