🔑 Developers & API

Agent-Safety API

Updated July 19, 2026

agentsmcpsafety

Agent-Safety API

Before an AI agent logs in, pays, or uploads, it can request a policy decision:

curl -X POST https://api.trustiry.com/agent/check \
  -H "Content-Type: application/json" \
  -d '{"url":"https://amaz0n-secure-login.example","action":"payment"}'

The response includes safe_to_proceed, blocked_actions, reason codes, evidence time, expiry, unknowns, and an official alternative when one is known. Sensitive actions fail closed when current relationship evidence is missing.