The Ethereum Foundation reports that its Protocol Security team is using coordinated AI agents to proactively test Ethereum’s critical network infrastructure for vulnerabilities. The initiative aims to shift security work toward identifying and validating real issues earlier in the process rather than relying only on traditional bug-finding workflows. According to the Foundation and related reporting, the AI systems are used to scan and analyze components of the network to surface potential security problems. The Block reports that while the AI agents do identify bugs, many initial findings are later determined to be false positives, indicating that the team is validating results rather than treating every alert as a confirmed flaw. Decrypt similarly describes the effort as using AI agents to find vulnerabilities before attackers do, with an emphasis on proving which issues are actually exploitable or legitimate. Together, the sources indicate that the Foundation’s approach involves automated testing plus follow-up verification to determine which suspected vulnerabilities are genuine. The reporting characterizes the work as part of an ongoing security process to improve the reliability and safety of Ethereum network infrastructure.