Woofun AI reports that the Ethereum Foundation Protocol Security team deployed specialized AI agents to conduct red team testing on ETH network infrastructure. These agents, assigned roles including reconnaissance and verification, identified a remotely triggerable panic vulnerability within the libp2p gossipsub peer-to-peer layer used by consensus clients. The flaw was patched and published as CVE-2026-34219. The foundation noted that while AI expands testing scope beyond manual review, it necessitates rigorous human judgment to evaluate findings.