构建自动化平台，系统评估网页智能体安全风险。
WebTrap Park: An Automated Platform for Systematic Security Evaluation of Web Agents
- 将1226个安全威胁转化为可执行测试任务，直接观察智能体与真实网页交互。
- 发现不同智能体框架存在明显安全差异，架构设计影响远超底层模型。
- 无需修改智能体即可评估，适合研究人员和开发者做安全验证。
网页智能体在真实网络环境中执行复杂任务日益普遍，但其安全评估仍零散且难以标准化。我们提出 WebTrap Park，一个通过直接观测智能体与实时网页交互实现系统性安全评估的自动化平台。该平台将三大类主要安全风险转化为1,226个可执行评估任务，支持无需修改智能体的动作级评估。实验结果揭示了不同智能体框架间显著的安全差异，凸显了智能体架构的重要性超越底层模型。WebTrap Park 已公开，网址为 https://security.fudan.edu.cn/webagent，为可复现的网页智能体安全评估提供可扩展基础。
原文摘要 · Abstract (English)
Web Agents are increasingly deployed to perform complex tasks in real web environments, yet their security evaluation remains fragmented and difficult to standardize. We present WebTrap Park, an automated platform for systematic security evaluation of Web Agents through direct observation of their concrete interactions with live web pages. WebTrap Park instantiates three major sources of security risk into 1,226 executable evaluation tasks and enables action based assessment without requiring agent modification. Our results reveal clear security differences across agent frameworks, highlighting the importance of agent architecture beyond the underlying model. WebTrap Park is publicly accessible at https://security.fudan.edu.cn/webagent and provides a scalable foundation for reproducible Web Agent security evaluation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。