Ethereum Foundation Set AI Agents Loose on Its Code: Here's What They Actually Found

The Ethereum Foundation’s protocol security team ran coordinated artificial intelligence (AI) agents against the code Ethereum depends on, surfacing at least one remotely exploitable bug along with a flood of convincing false positives that humans had to untangle.

The experiment, detailed in a blog post published July 9 by Nikos Baxevanis of the foundation’s protocol security team, was titled “The triage is the product.” The findings drew wide attention as the most flagged issues turned out to be false positives, even though real bugs were in the mix.

The headline discovery is real enough: the agents helped surface a remotely-triggerable panic in gossipsub, part of the libp2p peer-to-peer networking layer that Ethereum consensus clients run on. The flaw was fixed and disclosed as CVE-2026-34219, a bug that, if found first by an attacker, could have disrupted nodes across the network.

The surprise, the foundation wrote, was not that AI agents could find bugs but “how little of the work went into finding them, and how much went into telling the real bugs from the ones that just looked real.” The team catalogued the recurring shapes of those imposters: crashes that only occur in debug builds and never in production, reproducers that rely on unreachable internal values no attacker could supply, and formal-verification proofs that are technically true but so unconstrained they demonstrate nothing.

The foundation’s answer was a hard evidentiary standard: “reproducible or it didn’t happen.” Every candidate finding must ship with a self-contained artifact that reproduces the failure against the actual code, independent of how confident the reporting agent claims to be. Agents are viewed as hypothesis generators—search tools, not decision-makers—organized into recon, hunting, gap-filling, and validation stages, with humans making the final call.

The post offered a rare benchmark for how well the current generation of tools performs. A property-based testing agent generated roughly 1,000 candidate findings, and after expert review, about 86% of its top-tier recommendations survived scrutiny. This is strong for a machine, but a rate that still demands a human filter before anything touches production code.

While the tools are clearly finding real vulnerabilities in critical infrastructure, the workload has not disappeared but simply moved downstream to triage, where experienced engineers separate signal from simulation. For a network securing hundreds of billions of dollars, that filter is important.

The foundation is now pushing the work forward rather than treating it as a one-off. Its Ecosystem Support Program is funding a dedicated grant round for AI-powered protocol security, covering research, auditing, and vulnerability detection.

Source: https://news.bitcoin.com/ethereum-foundation-ai-agents-bugs-triage/