OpenAI代理在Hugging Face渗透中尝试调用外部模型
wtf OpenAI’s rogue agents tried to get other AI models (!) to help them during t…
Agent越权调用外部模型与安全机制突破是前沿热点，涉及具体技术路径，值得安全与Agent开发者关注。
wtf OpenAI’s rogue agents tried to get other AI models (!) to help them during the Hugging Face hack, according to a new report covered by the NYT.
据《纽约时报》报道，OpenAI 的失控智能体在 Hugging Face 被黑期间，曾试图让其他 AI 模型（！）协助它们。
They tried using an image model to solve CAPTCHAs. They also tried contacting Claude Haiku, DeepSeek, Kimi and Qwen.
它们尝试使用图像模型来破解验证码。此外，它们还试图联系 Claude Haiku、DeepSeek、Kimi 和 Qwen。
One agent collected exposed access keys in a list called “LOOT.” It ranked the keys and picked the five best to share with other agents.
其中一个智能体将泄露的访问密钥收集在一个名为“LOOT”的列表中，并对这些密钥进行排名，挑选出最好的五个与其他智能体共享。
They used link shorteners and screenshot services to get around internet restrictions, run code and send data back as images.
它们使用链接缩短器和截图服务来绕过互联网限制，运行代码并将数据以图片形式传回。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力