中文
相关论文

相关论文: MCPTox: A Benchmark for Tool Poisoning Attack on R…

200 篇论文

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…

密码学与安全 · 计算机科学 2026-03-26 Yulin Shen , Xudong Pan , Geng Hong , Min Yang

The proliferation of agentic AI coding assistants, including Claude Code, GitHub Copilot, Cursor, and emerging skill-based architectures, has fundamentally transformed software development workflows. These systems leverage Large Language…

密码学与安全 · 计算机科学 2026-01-27 Narek Maloyan , Dmitry Namiot

Large Language Model (LLM) agents are autonomous systems powered by LLMs, capable of reasoning and planning to solve problems by leveraging a set of tools. However, the integration of multi-tool capabilities in LLM agents introduces…

密码学与安全 · 计算机科学 2025-12-04 Zichuan Li , Jian Cui , Xiaojing Liao , Luyi Xing

LLM agents are beginning to invoke industrial asset-management tools through the Model Context Protocol (MCP), yet whether they can act reliably on this substrate for safety-critical \emph{Prognostics and Health Management (PHM)} is…

人工智能 · 计算机科学 2026-05-12 Tianjun Feng , Yunfeng Chen , Chun-Yi Tsai , Yihan Sun , Ayan Das , Kaoutar El Maghraoui , Shuxin Lin , Dhaval Patel

Autonomous AI agents powered by large language models (LLMs) with structured function-calling interfaces enable real-time data retrieval, computation, and multi-step orchestration. However, the rapid growth of plugins, connectors, and…

Large language models (LLMs) are increasingly deployed in agentic systems, where a fundamental task is mapping user intents to relevant external tools. Errors in tool selection can have severe outcomes, such as unauthorized data access,…

密码学与安全 · 计算机科学 2026-05-14 Jehyeok Yeon , Isha Chaudhary , Gagandeep Singh

The Model Context Protocol (MCP) has emerged as a widely adopted mechanism for connecting large language models to external tools and resources. While MCP promises seamless extensibility and rich integrations, it also introduces a…

密码学与安全 · 计算机科学 2025-07-10 Zhihao Li , Kun Li , Boyang Ma , Minghui Xu , Yue Zhang , Xiuzhen Cheng

The Model Context Protocol (MCP) is rapidly emerging as the middleware for LLM-based applications, offering a standardized interface for tool integration. However, its built-in security mechanisms are minimal: while schemas and declarations…

密码学与安全 · 计算机科学 2025-12-04 Biwei Yan , Yue Zhang , Minghui Xu , Hao Wu , Yechao Zhang , Kun Li , Guoming Zhang , Xiuzhen Cheng

Recent work has embodied LLMs as agents, allowing them to access tools, perform actions, and interact with external content (e.g., emails or websites). However, external content introduces the risk of indirect prompt injection (IPI)…

计算与语言 · 计算机科学 2024-08-06 Qiusi Zhan , Zhixiang Liang , Zifan Ying , Daniel Kang

AI-powered coding assistant tools have revolutionized the software engineering ecosystem. However, prior work has demonstrated that these tools are vulnerable to poisoning attacks. In a poisoning attack, an attacker intentionally injects…

密码学与安全 · 计算机科学 2023-12-12 Sanghak Oh , Kiho Lee , Seonhye Park , Doowon Kim , Hyoungshick Kim

The Model Context Protocol (MCP), introduced by Anthropic, provides a standardized framework for artificial intelligence (AI) systems to interact with external data sources and tools in real-time. While MCP offers significant advantages for…

密码学与安全 · 计算机科学 2025-05-06 Vineeth Sai Narajala , Idan Habler

We introduce ETOM, a five-level benchmark for evaluating multi-hop, end-to-end tool orchestration by LLM agents within a hierarchical Model-Context Protocol (MCP) ecosystem. Existing benchmarks often assess tools in isolation, overlooking…

人工智能 · 计算机科学 2026-01-21 Jia-Kai Dong , I-Wei Huang , Chun-Tin Wu , Yi-Tien Tsai

Modern AI agents execute real-world side effects through tool calls such as file operations, shell commands, HTTP requests, and database queries. A single unsafe action, including accidental deletion, credential exposure, or data…

人工智能 · 计算机科学 2026-05-07 Chenglin Yang

The Model Context Protocol (MCP) has recently emerged as a standardized interface for connecting language models with external tools and data. As the ecosystem rapidly expands, the lack of a structured, comprehensive view of existing MCP…

密码学与安全 · 计算机科学 2025-07-01 Zhiwei Lin , Bonan Ruan , Jiahao Liu , Weibo Zhao

The Model Context Protocol (MCP) has become a widely adopted interface for LLM agents to invoke external tools, yet learned monitoring of MCP tool-call traffic remains underexplored. In this article, the proposed detector is presented as an…

密码学与安全 · 计算机科学 2026-05-25 Sultan Zavrak

With advances in decision-making and reasoning capabilities, multimodal agents show strong potential in computer application scenarios. Past evaluations have mainly assessed GUI interaction skills, while tool invocation abilities, such as…

计算机视觉与模式识别 · 计算机科学 2025-11-12 Hongrui Jia , Jitong Liao , Xi Zhang , Haiyang Xu , Tianbao Xie , Chaoya Jiang , Ming Yan , Si Liu , Wei Ye , Fei Huang

AI agents increasingly call external tools (file system, network, APIs) through the Model Context Protocol (MCP). These tool calls are the agent's syscalls -- privileged operations with side effects on shared state -- yet today's safety…

密码学与安全 · 计算机科学 2026-04-21 Daeyeon Son

This paper introduces \textbf{FinMCP-Bench}, a novel benchmark for evaluating large language models (LLMs) in solving real-world financial problems through tool invocation of financial model context protocols. FinMCP-Bench contains 613…

人工智能 · 计算机科学 2026-03-27 Jie Zhu , Yimin Tian , Boyang Li , Kehao Wu , Zhongzhi Liang , Junhui Li , Xianyin Zhang , Lifan Guo , Feng Chen , Yong Liu , Chi Zhang

Today's AI agents are built on large language models (LLMs) equipped with tools to access and modify external environments, such as corporate file systems, API-accessible platforms and websites. AI agents offer the promise of automating…

计算机与社会 · 计算机科学 2026-03-26 Merlin Stein

Personal AI agents like OpenClaw run with elevated privileges on users' local machines, where a single successful prompt injection can leak credentials, redirect financial transactions, or destroy files. This threat goes well beyond…

人工智能 · 计算机科学 2026-04-07 Bowen Wei , Yunbei Zhang , Jinhao Pan , Kai Mei , Xiao Wang , Jihun Hamm , Ziwei Zhu , Yingqiang Ge