AI 导读
覆盖聊天、语音、浏览器和编程智能体，仅需一个端点、URL 或电话号码作为输入。 Fabraix 报告在 AgentHarm 上达到 78% 的攻击成功率，并在数十家财富 500 强企业的智能体中发现可利用漏洞。 预约演示 👀 https://t.co/vMN8bDPcMz
正文
It covers chat, voice, browser, and coding agents and requires only an endpoint, URL, or phone number as input.
Fabraix reports a 78% attack success rate on AgentHarm and exploitable findings in agents at dozens of Fortune 500 companies.
Book a demo 👀
https://t.co/vMN8bDPcMz
来源：@testingcatalog · x.com