<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>协调披露 :: 标签 :: x7peeps</title><link>https://x7peeps.com/tags/%E5%8D%8F%E8%B0%83%E6%8A%AB%E9%9C%B2/index.html</link><description/><generator>Hugo</generator><language>zh-CN</language><lastBuildDate>Thu, 29 Oct 2026 12:30:00 +0800</lastBuildDate><atom:link href="https://x7peeps.com/tags/%E5%8D%8F%E8%B0%83%E6%8A%AB%E9%9C%B2/index.xml" rel="self" type="application/rss+xml"/><item><title>模型自己找出五条攻击链：Anthropic 红队报告给防守方留下了什么</title><link>https://x7peeps.com/AI/08-%E5%AE%89%E5%85%A8%E4%B8%8EAI%E8%9E%8D%E5%90%88/%E6%A8%A1%E5%9E%8B%E8%87%AA%E5%B7%B1%E6%89%BE%E5%87%BA%E4%BA%94%E6%9D%A1%E6%94%BB%E5%87%BB%E9%93%BE-Anthropic%E7%BA%A2%E9%98%9F%E6%8A%A5%E5%91%8A%E7%BB%99%E9%98%B2%E5%AE%88%E6%96%B9%E7%95%99%E4%B8%8B%E4%BA%86%E4%BB%80%E4%B9%88/index.html</link><pubDate>Thu, 29 Oct 2026 12:30:00 +0800</pubDate><guid>https://x7peeps.com/AI/08-%E5%AE%89%E5%85%A8%E4%B8%8EAI%E8%9E%8D%E5%90%88/%E6%A8%A1%E5%9E%8B%E8%87%AA%E5%B7%B1%E6%89%BE%E5%87%BA%E4%BA%94%E6%9D%A1%E6%94%BB%E5%87%BB%E9%93%BE-Anthropic%E7%BA%A2%E9%98%9F%E6%8A%A5%E5%91%8A%E7%BB%99%E9%98%B2%E5%AE%88%E6%96%B9%E7%95%99%E4%B8%8B%E4%BA%86%E4%BB%80%E4%B9%88/index.html</guid><description>10 月 27 日，Anthropic 的研究团队发了一篇博客：他们用未发布的内部研究模型（文中称 Claude），在 Chrome、Safari、Android、Windows、Linux 五个真实目标上构造出五条完整的攻击链，后果涵盖一键远程代码执行、零点击 RCE、锁屏绕过和远程拒绝服务——报告原话是"little to no user interaction"下达成 critical impacts。所有漏洞都已按协调披露流程上报并修补。文章署名作者包括 Nicholas Carlini、David Dworken、Xinyang Ge、Ben Morris、Milad Nasr、Chris Rohlf、Luna Tong——这份名单本身就是安全研究的分量证明。</description></item></channel></rss>