WuShuo has learned that Anthropic's latest research shows that AI agents possess significant on-chain attack capabilities: in simulated tests on real hacked smart contracts from 2020–2025, Claude Opus 4.5, Sonnet 4.5, and GPT-5 collectively replicated vulnerabilities worth approximately $4.6 million. When scanning 2,849 contracts with no known vulnerabilities, the two models also discovered 2 brand-new zero-day vulnerabilities and successfully simulated profits. The research points out that the on-chain attack profits of AI have doubled approximately every 1.3 months over the past year, and they are now technically fully capable of autonomous and profitable exploitation of vulnerabilities. https://www.(wublock123.com)/index.php?m=content&c=index&a=show&catid=6&id=52788