在上一篇文章 《罗福莉访谈之后:Vibe Coding → Vibe Working → Vibe Forking》https://mp.weixin.qq.com/s/D4bAcI3TN4b_cEhk-aftNQ 提
在上一篇文章 《罗福莉访谈之后:Vibe Coding → Vibe Working → Vibe Forking》https://mp.weixin.qq.com/s/D4bAcI3TN4b_cEhk-aftNQ 提到的另外一个“ah moment”其实就是指漏洞挖掘的模型分水岭出现了~ 今天看到两篇文章
介绍微软用他们的智能体 MDASH(multi-model agentic scanning harness),编排 100+ specialized AI agents,并称其在公开 CyberGym benchmark 上拿到 88.45% 并且在他们内部发现了多个tcpip.sys的RCE:
https://www.geekwire.com/2026/microsofts-multi-agent-ai-system-tops-anthropics-mythos-on-cybersecurity-benchmark/
https://www.microsoft.com/en-us/security/blog/2026/05/12/defense-at-ai-speed-microsofts-new-multi-model-agentic-security-system-tops-leading-industry-benchmark/