# Aztec Discloses a Critical Flaw That Lets a Forged Proof Pass Verification

Core contributors found the defect on 27 July through internal artificial-intelligence-assisted auditing, and say funds, applications and contract state on the V5 alpha network are exposed.

- Published: 2026-08-13T05:53:29.271Z
- Canonical: https://polylog.news/crypto/2026-08-13/aztec-discloses-a-critical-flaw-that-lets-a-forged-proof-pas
- Publisher: Polylog (Crypto desk)
- Section: crypto
- Sources: [Aztec Network](https://aztec.network/blog/alpha-v5-proving-system-vulnerability), [Aztec Network (Alpha V5)](https://aztec.network/alpha-v5), [The Defiant](https://thedefiant.io/news/defi/aztec-v4-withdraw-june-25-v5-upgrade-security-vulnerability)

Aztec, the privacy-focused layer-2 network on Ethereum that runs private smart contracts, [disclosed a critical vulnerability in the proving system](https://aztec.network/blog/alpha-v5-proving-system-vulnerability) behind its Alpha V5 release. Contributors identified the defect on 27 July through internal auditing assisted by artificial-intelligence tooling. Aztec states that an attacker may be able to construct a proof that passes verification for a transaction the network should reject, and that funds, applications and contract state on V5 should be treated as exposed until incident response finishes and operators complete the required network actions.

This is a soundness failure, not a bug in an application contract. A zero-knowledge rollup replaces re-execution with a mathematical proof, so if the verifier can be convinced by an invalid proof, the protocol has no additional check behind it. Validators cannot catch the problem by re-running transactions, because the design deliberately removed that step. That is what separates proving-system defects from ordinary smart-contract exploits and why teams disclose them on a schedule tied to upgrades rather than immediately.

It is also the second such disclosure in the same code lineage this year. Aztec previously [told V4 users to withdraw ahead of a June deadline](https://thedefiant.io/news/defi/aztec-v4-withdraw-june-25-v5-upgrade-security-vulnerability) after a critical flaw the team said affected the proving system as a whole. V5, released with roughly halved private-transaction proving costs, remains [labeled alpha software](https://aztec.network/alpha-v5), and Aztec has said this is the phase in which findings of this kind surface. No exploitation has been reported.

## What this means

Every zero-knowledge rollup asks users to trust a verifier contract and the circuits behind it, and a soundness bug converts that trust into an unbacked claim on the bridge holding the deposits. The exposed parties are depositors on the privacy network and any application settling through it. Two paths follow from here. Contributors ship fixes and the disclosure becomes evidence that pre-mainnet auditing works, or a similar defect surfaces after real capital arrives, which would push serious users toward multiple independent proof systems and emergency withdrawal mechanisms rather than a single prover.

## What to watch

- Whether Aztec publishes the promised security roadmap and names the specific circuit or verifier component at fault, which tells other teams using the same proving stack whether they inherit the bug.
- Whether other zero-knowledge rollups adopt second proof systems or manual emergency withdrawal mechanisms, a direct cost that shows how seriously operators take verifier risk.
- How much of the disclosure pipeline now comes from artificial-intelligence-assisted review rather than external bounty hunters, which changes who finds these defects first.