Zenity 沙箱检测数千公开 Agent skills 发现窃密活动

Zenity 在沙箱中引爆数千个公开 Agent 技能，发现一场安装量达 170 万次的窃取凭据活动，一个排名前 150 的技能数月未被检测到，还有智能体被改造成恶意软件投放器。这些技能并非被劫持，而是从一开始就被构建成这样。Zenity 同时推出免费开放的 AI Total，并附上博客链接 https://zenity.io/blog/introducing-ai-total。

Zenity Labs released a free service called AI Total that uses a Detonation Chamber to run Agent Skills in a sandbox, recording their behavior and comparing it against claims. Before release, it detonated thousands of public Skills and found dozens of malicious samples, one of which—a credential-stealing campaign distributed via Vercel skills.sh—reached approximately 1.7 million installs; over 30% of malicious Skills instructed agents like Claude Code and OpenClaw to download and run attacker files.