Aztec Discloses a Proving-System Flaw That Could Let Forged Proofs Pass on Its Alpha V5 Network

Aztec, a privacy-focused layer-2 network on Ethereum, disclosed a critical vulnerability in the proving system of its Alpha V5 network. Contributors say they found the flaw on July 27 through internal AI-assisted auditing. An attacker could build a proof that passes verification for a transaction the network should reject. If the network accepted such a transaction, it could change state in ways V5's rules are meant to forbid.

Aztec told users to treat V5 funds, applications and contract state as exposed to a protocol-level failure. That warning stands until contributors finish their incident response and node operators carry out the required network actions. Aztec also said it cannot tell whether anyone exploited the flaw. The system does not keep the information that would separate a properly accepted transaction from one accepted through the flawed path. Chain history therefore cannot show whether exploitation happened or how much was taken. The fix is planned for V6, with circuit changes that stop the network from accepting proofs tied to the affected proving system.

The flaw is hard to detect because of how Aztec works. On a transparent rollup, an outside observer can re-run transactions and catch an invalid state change. On Aztec, private functions run on the user's own device, and the network receives only a zero-knowledge proof and encrypted commitments. If a false proof verifies, observers have nothing to re-check. The disclosure was published well after the July discovery and tied to the next release. That timing follows the usual practice of keeping a bug private until a fix ships.

This is not the network's first proof-layer incident. Aztec previously published a critical vulnerability in Alpha v4. Meanwhile, the project keeps building its user base. It has launched an onchain version of the Dark Forest game and published an explainer on how gas works on the network.