Mitigating Taint-Style Vulnerabilities in MCP Servers via Security-Aware Tool Descriptions

arXiv:2607.07461 · cs.CR, cs.SE · Submitted 2026-07-08 · Read on arXiv

Yang Shi, Jiaheng Fu, Yihe Huang, Ruixiang Wu, Chengyao Sun, Kaifeng Huang

cs.CR, cs.SE

Submitted: 2026-07-08

Code: https://github.com/en/rest

Project page: https://openai.com/index/hello-gpt-4o

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Model Context Protocol — 18× in this paper
• Benchmark — 12× in this paper · explained in 2 episodes
• Model Context Protocol (MCP) — 6× in this paper · explained in 8 episodes
• Ablation — 3× in this paper · explained in 1 episode
• Adversarial Attacks — 3× in this paper · explained in 1 episode
• Large Language Models (LLMs) — 3× in this paper · explained in 6 episodes
• Ablation study — 2× in this paper · explained in 1 episode
• Generalizability — 2× in this paper · explained in 6 episodes
• National Vulnerability Database — 2× in this paper
• Prompt Engineering — 2× in this paper · explained in 4 episodes
• Robustness — 2× in this paper · explained in 5 episodes
• Adam — 1× in this paper · explained in 2 episodes
• Adversarial Prompting — 1× in this paper · explained in 1 episode
• Benchmarking — 1× in this paper · explained in 3 episodes
• Core — 1× in this paper · explained in 2 episodes
• IEEE Symposium on Security and Privacy — 1× in this paper
• Low-Resource Languages — 1× in this paper · explained in 4 episodes
• National Institute of Standards and Technology — 1× in this paper
• Server-Side Request Forgery — 1× in this paper

Sources

• Model Context Protocol (MCP) at First Glance: Studying the Security and Maintainability of MCP Servers
• Baseline Defenses for Adversarial Attacks Against Aligned Language Models
• AgentBound: Securing Execution Boundaries of AI Agents
• Improved Techniques for Optimization-Based Jailbreaking on Large Language Models
• MCPToolBench++: A Large Scale AI Agent Model Context Protocol MCP Tool Use Benchmark
• Certifying LLM Safety against Adversarial Prompting
• MCP Guardian: A Security-First Layer for Safeguarding MCP-Based AI System
• Don't believe everything you read: Understanding and Measuring MCP Behavior under Misleading Tool Descriptions
• A Large-Scale Evolvable Dataset for Model Context Protocol Ecosystem and Security Analysis
• AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models
• Jailbreaking ChatGPT via Prompt Engineering: An Empirical Study
• MCPZoo: A Large-Scale Dataset of Runnable Model Context Protocol Servers for AI Agent
• LLM Self Defense: By Self Examination, LLMs Know They Are Being Tricked
• SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks
• MCPMark: A Benchmark for Stress-Testing Realistic and Comprehensive MCP Use
• MCP-Guard: A Multi-Stage Defense-in-Depth Framework for Securing Model Context Protocol in Agentic AI
• Beyond the Protocol: Unveiling Attack Vectors in the Model Context Protocol (MCP) Ecosystem
• Help or Hurdle? Rethinking Model Context Protocol-Augmented Large Language Models
• The Dark Side of Trust: Authority Citation-Driven Jailbreak Attacks on Large Language Models
• MPMA: Preference Manipulation Attack Against Model Context Protocol

Related papers

• SoK: AI-Augmented Binary Reversing
• Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
• Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Efficient Fuzzy PSI under One-Sided Assumptions
• Sealing the Audit-Runtime Gap for LLM Skills
• Token Composition: A Graph Based on EVM Logs