AI agents are starting to work with other agents — across teams, companies, and platforms. But there's no standard way to verify who's on the other end. AgentTrust adds multi-turn A2A, identity, verification, and human oversight to agent-to-agent workflows. Built on Google's A2A protocol. → Ed25519 signed messages (tamper-proof, non-repudiable) → Human-in-the-loop escalation rules → 13 tools via MCP server — one install → InjectionGuard for untrusted inputs