Analyzes, diagnoses, and optimizes prompts to improve LLM responses by detecting issues and recommending concrete edits.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.