I Sent a WhatsApp Message to an AI Agent. It Ran My Code on the Host.
中文摘要
OpenClaw AI 助手被发现存在三个高危漏洞，可被利用进行远程代码执行，使其沦为远程访问工具。
English Summary
OpenClaw AI agent has three high-severity vulnerabilities allowing remote code execution, potentially turning assistants into remote access tools.
原文节选
Three high-severity vulnerabilities in OpenClaw turn your personal AI assistant into a remote access tool — and the "state-of-the-art"… Continue reading on Medium »