AC2 Protocol: The missing security layer for AI agents

Post by josephcecala » Thu, Aug 27, 2026, 2:42 PM UTC

Miyamura80 wrote: post #3

josephcecala wrote: post #2

AC2 is an open standard that puts users back in control of AI-driven signing operations, providing verifiable proof of intent and credential isolation. The problem: a compromised agent runtime (e.g. a malicious plugin dependency) can leak everything injected into it like API keys or session tokens, and there's no way to prove whether a human actually approved what happened, since chat-based "approvals" are just messa…

This is cool, we've built similar things, I'm a bit uneasy with crypto stuff involved, especially with algorand But FIDO2 passkey signature I absolutely agree should be the way to go. Just need more open standard without tokens involved

Post by josephcecala » Thu, Aug 27, 2026, 2:45 PM UTC

embedding-shape wrote: post #10

josephcecala wrote: post #8

Earlier quoted context omitted.

Good questions, the core doesn't need a blockchain at all. AC2 is just an approval/signing layer (DIDComm + WebAuthn/FIDO2 + WebRTC), so most use cases like API access, git commits, and client comms never touch a chain. Blockchain only enters when what's being signed is a transaction, like an x402 payment. For that case we use Algorand, but AC2 is chain-agnostic, meaning any chain could plug in. Algorand fit because…

Judging by my own initial reaction, and other commentators here too, if this tech doesn't actually need any cryptocurrency to be useful, then probably you shouldn't mention that at all on the landing page, but let the people who need that stuff find it on some other pages. You'll fight a huge deluge of knee-jerk reactions otherwise, as you can tell :) Regardless, thanks for expanding and explaining, and good luck to…

Post by chrisjj » Thu, Aug 27, 2026, 2:55 PM UTC

Post by josephcecala » Thu, Aug 27, 2026, 3:27 PM UTC

otabdeveloper4 wrote: post #16

josephcecala wrote: post #2

AC2 is an open standard that puts users back in control of AI-driven signing operations, providing verifiable proof of intent and credential isolation. The problem: a compromised agent runtime (e.g. a malicious plugin dependency) can leak everything injected into it like API keys or session tokens, and there's no way to prove whether a human actually approved what happened, since chat-based "approvals" are just messa…

Do I understand this right? As long as you can prove that a real human "approved" sharing the God-mode long-lived token with the keys to kingdom with an LLM, it's all good and we're secure? This is the state of the art of computer security?

Post by semiquaver » Thu, Aug 27, 2026, 7:10 PM UTC

josephcecala wrote: post #11

Miyamura80 wrote: post #3

Earlier quoted context omitted.

This is cool, we've built similar things, I'm a bit uneasy with crypto stuff involved, especially with algorand But FIDO2 passkey signature I absolutely agree should be the way to go. Just need more open standard without tokens involved

Fair, and to be clear, you can use AC2 today with zero crypto exposure. The whole protocol for API access, git commits, client comms uses FIDO2 signing, DIDComm, and WebRTC. No wallet, no chain, no token needed. Algorand only shows up in the reference implementation for the payment use case (x402), since something has to settle an actual transfer. That's opt-in, not required. Skip payments and you never touch it.