Skip to content
Tech News
← Back to articles

AC2 Protocol: The missing security layer for AI agents

read original more articles
Why This Matters

The AC2 protocol introduces a robust security layer for AI agents, enabling secure, verifiable interactions without reliance on central relays or blockchains. Its integration of DIDComm v2.0, passkey authentication, and real-time streaming enhances trust and interoperability across various workflows, from code deployment to payments. This development is significant for the tech industry as it promotes more secure, transparent, and decentralized AI operations, benefiting both developers and consumers.

Key Takeaways

What you can prove

Upgrade your existing frameworks with one plugin and one command line. AC2 uses DIDComm v2.0 message formats for interoperability, passkey-based authentication via Liquid Auth (built on FIDO2/WebAuthn), and supports real-time voice and text streaming alongside signing delegation. It requires no central message relay and no blockchain to operate. AC2 supports any type of agentic workflow – you can use it for any custom schema you define via JSON-LD.

Here are some examples of what you can do with AC2:

Code deploys

Signed commits with hardware-backed proof that a human exercised final authority over the change set before the merge.

Client communications

A signed approval on the exact message body sent by the agent, proof a human signed off before anything reached the outbox.

API access

The agent presents exact request parameters for approval before execution, the signature proves what was authorized, by whom, and when.

x402 payments

... continue reading