arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2512.04129cs.CR
收藏

推倒多米诺骨牌:面向基于大语言模型的多智能体系统的拓扑感知多跳攻击

Tipping the Dominos: Topology-Aware Multi-Hop Attacks on LLM-Based Multi-Agent Systems

Nanyang Technological University(南洋理工大学);Wuhan University(武汉大学);Beijing Institute of Technology(北京理工大学)

Ruichao Liang, Le Yin, Jing Chen, Cong Wu, Xiaoyu Zhang, Huangpeng Gu, Zijian Zhang, Yang Liu

AI总结:

针对基于大语言模型的多智能体系统,提出拓扑感知多跳攻击方法,揭示其内在安全漏洞并提出基于拓扑信任的防御框架。

摘要:

基于大语言模型的多智能体系统(MASs)凭借其涌现的协调和问题解决能力重塑了数字景观。然而,当前对MASs的安全评估仍局限于有限的攻击场景,使其安全问题不明确且可能被低估。为填补这一空白,我们提出了TOMA,一种面向MASs的拓扑感知多跳攻击方案。通过优化MAS拓扑内污染的传播并控制环境起源的对抗性负载的多跳扩散,TOMA揭示了新的有效攻击向量,而无需特权访问或直接智能体操控。实验显示,在三种最先进的MAS架构:Magentic-One、LangManus和OWL,以及五种代表性拓扑中,攻击成功率从40%到78%不等,揭示了可能被现有研究忽视的内在MAS漏洞。受这些发现启发,我们提出了一种基于拓扑信任的概念性防御框架,并原型实验显示其在阻止94.8%的适应性和复合攻击方面具有有效性。

英文摘要:

LLM-based multi-agent systems (MASs) have reshaped the digital landscape with their emergent coordination and problem-solving capabilities. However, current security evaluations of MASs are still confined to limited attack scenarios, leaving their security issues unclear and likely underestimated. To fill this gap, we propose TOMA, a topology-aware multi-hop attack scheme targeting MASs. By optimizing the propagation of contamination within the MAS topology and controlling the multi-hop diffusion of adversarial payloads originating from the environment, TOMA unveils new and effective attack vectors without requiring privileged access or direct agent manipulation. Experiments demonstrate attack success rates ranging from 40% to 78% across three state-of-the-art MAS architectures: \textsc{Magentic-One}, \textsc{LangManus}, and \textsc{OWL}, and five representative topologies, revealing intrinsic MAS vulnerabilities that may be overlooked by existing research. Inspired by these findings, we propose a conceptual defense framework based on topology trust, and prototype experiments show its effectiveness in blocking 94.8% of adaptive and composite attacks.

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。

↑