Ethereum Foundation Uses AI Agents to Test ETH Network and Discover Vulnerabilities

The Ethereum Foundation's Protocol Security team announced in a blog post on Thursday that they have deployed AI agents to test the software relied upon by Ethereum, searching for vulnerabilities in the cryptocurrency system, protocol code, and smart contracts. The AI agents discovered a remotely triggerable panic issue present in the libp2p gossipsub of the Ethereum consensus client, which has been fixed and disclosed on GitHub as CVE-2026-34219. Researchers stated that the AI agents were organized into roles such as reconnaissance, searching, patching, and verification to identify potential attack paths, reproduce failures, and validate their applicability. The Ethereum Foundation emphasized that AI does not replace security researchers but changes the way they work, enabling the team to cover a broader scope, while researchers still need to exercise caution in judging a large number of seemingly credible conclusions.