Security researchers have disclosed a high-severity vulnerability dubbed "ClawJacked" in the popular AI agent OpenClaw that ...

A serious vulnerability in the open-source AI agent OpenClaw made it possible for arbitrary websites to take complete control ...

PANews 3月2日消息，GoPlus中文社区发布预警，OpenClaw ...

North Korean-linked campaign publishes 26 malicious npm packages hiding C2 in Pastebin, deploying credential stealers & RAT ...

Oasis安全研究人员在OpenClaw中发现了一个关键的零交互漏洞。作为史上增长最快的开源AI Agent框架之一，该漏洞允许任何恶意网站无需插件、扩展或用户操作即可静默获取开发者AI Agent的完全控制权。

OpenClaw修复了一个高危安全漏洞ClawJacked，该漏洞可能允许恶意网站通过WebSocket连接到本地运行的AI代理并获取控制权。攻击者可利用JavaScript暴力破解网关密码，注册为受信任设备，完全控制AI代理。此外，ClawHub平台发现71个恶意技能包，部分伪装成加密货币工具实施诈骗。微软建议将OpenClaw视为不可信代码执行环境，仅在完全隔离的环境中部署。

Oasis Security reveals how a new ClawJacked vulnerability could allow attackers to silently take over a victim’s OpenClaw ...