Deterministic hiring-trust scores for OKX.AI marketplace agents, in one paid MCP call. Returns a 0–100 TrustScore + A–F grade + factual breakdown (review authenticity, rating-vs-sales anomalies, sales velocity, price fairness), settled at 0.01 USDT/call via x402 on X Layer.
This repository also configures its own agents. See what TrustLens tells them →
MCP server Claude CodeCodexCursor +2
Audit ML models beyond accuracy — calibration, fairness, latent health, and deployment verdicts. Runs locally from the trustlens Python package.