Primary sourceIT之家
View sourceEvent ·Multi-source
白帽黑客借 Anthropic Claude 入侵 OpenAI ChatGPT 账户,获 6500 美元赏金
独立安全团队 Hacktron AI 利用 Anthropic Claude 4.8 特殊版本成功入侵 OpenAI 员工的 ChatGPT 账户,获取了该公司私有软件代码库文档。他们通过发现 Discourse 论坛的 CVE-2026-45788 漏洞,利用 Claude 编写攻击代码,最终获取了有效登录 token,并成功访问了 OpenAI 的 Monorepo 仓库。OpenAI 已确认并修复了该漏洞。
Current read
朋友,有个挺有意思的事,安全公司 Hacktron AI 用 Anthropic 的 Claude 模型入侵了 OpenAI 的 ChatGPT 账户,还拿了 6500 美元的赏金。他们发现 Discourse 论坛有个漏洞,然后让 Claude 写了攻击代码,最后拿到了 OpenAI 员工的登录 token,进了他们的代码库。
11 sources78° AI heatLast updated 2026/9/18 05:26:43
Evidence
11 sourcesRelated source 1宝玉
View sourceRelated source 2techcrunch
View sourceRelated source 3theverge
View sourceRelated source 4lmarena.ai
View sourceRelated source 5Decoder
View sourceRelated source 6Gary Marcus
View sourceRelated source 7Genk
View sourceRelated source 8AI Will
View sourceRelated source 9Rappler: Technology
View sourceRelated source 10Anthropic
View sourceConflict check
Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.