arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.20156cs.CRcs.LG
收藏

Skill-Inject: 评估代理对技能文件攻击的脆弱性

Skill-Inject: Measuring Agent Vulnerability to Skill File Attacks

Max Planck Institute for Intelligent Systems, ELLIS Institute Tübingen, Tübingen AI Center(马克斯·普朗克智能系统研究所、图宾根ELLIS研究所、图宾根人工智能中心)

David Schmotz, Luca Beurer-Kellner, Sahar Abdelnabi, Maksym Andriushchenko

AI总结:

SkillInject通过评估代理对技能文件攻击的脆弱性,揭示了当前LLM代理在安全性和合规性方面的严重问题,指出需采用上下文感知的授权框架来提升安全性。

摘要:

大型语言模型代理正在迅速发展,通过代码执行、工具和最近引入的代理技能功能进行驱动。技能允许用户通过专门的第三方代码、知识和指令扩展LLM应用。尽管这可以将代理的能力扩展到新领域,但这也创造了一个日益复杂的代理供应链,提供了新的提示注入攻击的表面。我们识别出基于技能的提示注入是一个重大威胁,并引入SkillInject,一个评估广泛使用的LLM代理对通过技能文件注入的易受攻击性的基准。SkillInject包含202个注入任务对,攻击范围从明显恶意的注入到隐蔽在看似合法指令中的微妙、上下文依赖的攻击。我们评估了前沿LLM在SkillInject上的表现,测量了安全方面的有害指令规避和实用性方面的合法指令合规性。我们的结果表明,当今的代理高度易受攻击,前沿模型的攻击成功率高达80%,经常执行极其有害的指令,包括数据外泄、破坏性行为和类似勒索软件的行为。它们还表明,这个问题无法通过模型扩展或简单的输入过滤解决,但需要基于上下文的授权框架来实现稳健的代理安全。我们的基准可在https://www.skill-inject.com/上获取。

英文摘要:

LLM agents are evolving rapidly, powered by code execution, tools, and the recently introduced agent skills feature. Skills allow users to extend LLM applications with specialized third-party code, knowledge, and instructions. Although this can extend agent capabilities to new domains, it creates an increasingly complex agent supply chain, offering new surfaces for prompt injection attacks. We identify skill-based prompt injection as a significant threat and introduce SkillInject, a benchmark evaluating the susceptibility of widely-used LLM agents to injections through skill files. SkillInject contains 202 injection-task pairs with attacks ranging from obviously malicious injections to subtle, context-dependent attacks hidden in otherwise legitimate instructions. We evaluate frontier LLMs on SkillInject, measuring both security in terms of harmful instruction avoidance and utility in terms of legitimate instruction compliance. Our results show that today's agents are highly vulnerable with up to 80% attack success rate with frontier models, often executing extremely harmful instructions including data exfiltration, destructive action, and ransomware-like behavior. They furthermore suggest that this problem will not be solved through model scaling or simple input filtering, but that robust agent security will require context-aware authorization frameworks. Our benchmark is available at https://www.skill-inject.com/.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑