Ethereum Foundation: An AI discovered a vulnerability that could cause validator nodes to go offline, but manual verification is still required to confirm whether it is real.

Deep Tide TechFlow update: On July 11, CoinDesk reported that the Ethereum Foundation recently disclosed that its security team tested the software running on Ethereum validator nodes using AI agents and successfully found a vulnerability that can be triggered remotely, causing the nodes to crash. However, researchers emphasized that despite the large volume of security reports generated by AI, manual review is still the key step to distinguish real vulnerabilities from false positives.

It is reported that the vulnerability discovered this time exists in the Ethereum network message propagation protocol gossipsub. An attacker can remotely trigger the node software to enter an abnormal computing state, causing the program to crash and shut down, which takes the validator node offline until the operator manually restarts it. The vulnerability has been fixed and registered under the identifier “CVE-2026-34219”.

Nikos Baxevanis, a member of the Ethereum Foundation protocol security team, said that what was truly surprising about this incident was not the ability of AI to find the vulnerability, but the fact that the team spent a large amount of time determining which vulnerabilities genuinely existed and which were merely seemingly credible “illusions.”