Show HN: Approve an AI agent's wire with Face ID,then watch a forged one fail

Category: security

Tags: trust-protocol, ai-agent, authorization, formal-verification, web-authn

Score: 8.3/10 (Innovation: 8, Technical: 9, Documentation: 9, Utility: 7)

EMILIA Protocol is a protocol-grade trust substrate that enforces cryptographic consent before high-risk AI agent actions, using Face ID (WebAuthn) signoff and offline-verifiable receipts. It combines formal verification (TLA+, Alloy), multiple independent reference implementations (JS, Python, Go), and a novel pre-action authorization flow called Handshake, making it an innovative and technically impressive solution for agent governance.

Target audience: backend devs, devops, security engineers

Repository: https://www.emiliaprotocol.ai/try · JavaScript · Apache-2.0 · 2 stars

View on Hacker News