Thoughtproof

x402 degraded

Multi-LLM reasoning verification before agent settlement. Returns ALLOW/HOLD/UNCERTAIN/DISSENT with confidence score and EdDSA-signed attestation. Prevents agents from acting on weak or faulty reasoning.