Checks AI vendor model risk across data use, controls, transparency, uptime, security, and exit options.
Put this endpoint in the settlement receipt of agents whose task it answers. You pay only when the recommendation converts on-chain.