Exchange Ranking
Top Exchanges
24h Volume Ranking
Popularity Ranking
Exchange BTC Balance
Proof of Reserves
Decentralized Exchanges
Funding Rate
Funding Heatmap
Liquidation Data
Max Pain
Long/Short Ratio
Whale L/S Ratio
Binance/Okex/Huobi L/S
Bitfinex Margin L/S
ETF Tracker
Solana ETF
XRP ETF
Hong Kong ETF
Bitcoin Treasuries
Crypto Reversal
Ethereum Reserves
HyperLiquid Wallet Analysis
Hyperliquid Whale Watch
Large Transactions
On-chain Movement
Bitcoin ROI
Stablecoin Market Cap
Options Analysis
News
Articles
Economic Calendar
Features
Wallet
Contract Calculator
Security
Collections
Watchlist
Following
ETH The Ethereum Foundation said that while Ethereum (ETH) was preparing for major protocol changes, collaborative AI agents discovered real software vulnerabilities during security testing.
Key Points:
The AI agent identified a real vulnerability, including a remotely triggered consensus client failure.
Researchers say automated systems expand test coverage, but can also produce convincing false positives.
Human security experts remain the key to verifying findings and judging their severity.
Ethereum AI testing
The foundation\'s protocol security team said its collaborative agent discovered real vulnerabilities while inspecting Ethereum code, including a remotely triggered consensus client point-to-point layer panic error. Developers have fixed the issue.
The researchers point out that the unexpected result was not that the agent discovered the vulnerability, but that the process of distinguishing valid discoveries from seemingly reasonable but actual errors was more laborious than discovering the vulnerability itself.
\"What\'s really surprising is that there is so little work required to find vulnerabilities and so much work is needed to distinguish real vulnerabilities from seemingly real fake vulnerabilities,\" the team wrote.
The foundation divides its workflow into reconnaissance, hunting, gap filling and verification roles, drawing on earlier research from Anthropic and Cloudflare. It said agents did a good job in associating specifications with code and raising possible root causes.
However, these systems sometimes view unreachable call chains as exploitable and exaggerate the severity of the vulnerability. The team said that human reviewers still need to test whether the reported vulnerabilities are authentic and assess their actual impact.
\"Agents allow us to cover much more scope than manual testing,\" the post read. \"In exchange, they require us to make more cautious judgments amid a large number of confidence-sounding assertions.\"
Ethereum Security Transformation
This experiment comes as the foundation narrows its focus to underlying development, cryptographic protection and emergency security fixes. In June, it also outlined plans to allocate more responsibilities to the broader Ethereum ecosystem.
The nonprofit organization has laid off 20% of its workforce, cut its annual budget by nearly half, and disbanded its team, including agreement support. New organizations, including Ethlabs and Ethereum Institutional, have been formed to provide additional research capabilities.
This broader reform is related to the proposed \"Streamlining Ethereum\" roadmap, which Vitalik Buterin has said could take four years to complete. The plan will affect multiple network components while simultaneously aiming to achieve faster performance, greater resilience and the ability to withstand future quantum computing threats.
The last such shift for Ethereum was a \"merger\", which transitioned the network from proof-of-work mining to proof-of-stake verification. Current security initiatives reflect the greater engineering burden imposed by another planned redesign.
