arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.24203cs.CRcs.AI
收藏

基于模型上下文协议的隐形威胁:通过树状自适应搜索生成隐蔽注入负载

Invisible Threats from Model Context Protocol: Generating Stealthy Injection Payload via Tree-based Adaptive Search

School of Computer Science, Fudan University(复旦大学计算机科学学院);Shanghai Innovation Institute(上海创新研究院);IEEE Publication Technology Group(IEEE出版技术组)

Yulin Shen, Xudong Pan, Geng Hong, Min Yang

AI总结:

本文提出TIP攻击方法,通过树状结构搜索生成隐蔽负载,有效攻击MCP增强代理,实验显示其在未防御环境下攻击成功率超95%,且优于现有方法。

摘要:

近期模型上下文协议(MCP)的进步使大语言模型(LLMs)能够轻松调用外部工具,从而产生新的强大且工具增强的代理。然而,这种能力也引入了未被探索的攻击面,即恶意操控工具响应。现有针对MCP的间接提示注入技术存在部署成本高、语义不连贯或白盒要求重的问题,且易被新提出的防御措施检测。本文提出Tree structured Injection for Payloads (TIP),一种新的黑盒攻击,生成自然负载以可靠控制MCP增强代理,即使在防御下也能有效。技术上,将负载生成视为树状结构搜索问题,并通过我们的粗到细优化框架指导搜索。为稳定学习并避免局部最优,引入路径感知反馈机制,仅将高质量历史轨迹反馈给攻击者模型。框架进一步通过显式条件化搜索于可观察的防御信号并动态重新分配探索预算,以增强对防御转换的抵抗力。在四个主流LLMs上的大量实验显示,TIP在无防御环境下攻击成功率超过95%,所需查询次数比现有自适应攻击少一个数量级。针对四个代表性防御方法,TIP保持超过50%的有效性,并显著优于现有最先进攻击。通过在现实世界MCP系统中实施攻击,我们的结果揭示了MCP部署中的隐形但实用的威胁向量。我们还讨论了潜在的缓解方法以解决这一关键安全缺口。

英文摘要:

Recent advances in the Model Context Protocol (MCP) have enabled large language models (LLMs) to invoke external tools with unprecedented ease. This creates a new class of powerful and tool augmented agents. Unfortunately, this capability also introduces an under explored attack surface, specifically the malicious manipulation of tool responses. Existing techniques for indirect prompt injection that target MCP suffer from high deployment costs, weak semantic coherence, or heavy white box requirements. Furthermore, they are often easily detected by recently proposed defenses. In this paper, we propose Tree structured Injection for Payloads (TIP), a novel black-box attack which generates natural payloads to reliably seize control of MCP enabled agents even under defense. Technically, We cast payload generation as a tree structured search problem and guide the search with an attacker LLM operating under our proposed coarse-to-fine optimization framework. To stabilize learning and avoid local optima, we introduce a path-aware feedback mechanism that surfaces only high quality historical trajectories to the attacker model. The framework is further hardened against defensive transformations by explicitly conditioning the search on observable defense signals and dynamically reallocating the exploration budget. Extensive experiments on four mainstream LLMs show that TIP attains over 95% attack success in undefended settings while requiring an order of magnitude fewer queries than prior adaptive attacks. Against four representative defense approaches, TIP preserves more than 50% effectiveness and significantly outperforms the state-of-the-art attacks. By implementing the attack on real world MCP systems, our results expose an invisible but practical threat vector in MCP deployments. We also discuss potential mitigation approaches to address this critical security gap.

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。

↑