Analyzes text to detect and flag prompt-injection attempts, helping secure AI prompts and agent interactions.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.