Scores how well an AI workload fits sovereign requirements like residency, sensitivity, latency, auditability, and control.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.