Score: 0

Securing the Model Context Protocol: Defending LLMs Against Tool Poisoning and Adversarial Attacks

Published: December 6, 2025 | arXiv ID: 2512.06556v1

By: Saeid Jamshidi , Kawser Wazed Nafi , Arghavan Moradi Dakhel and more

Potential Business Impact:

Secures AI tools from hidden, dangerous instructions.

Business Areas:
Semantic Search Internet Services

The Model Context Protocol (MCP) enables Large Language Models to integrate external tools through structured descriptors, increasing autonomy in decision-making, task execution, and multi-agent workflows. However, this autonomy creates a largely overlooked security gap. Existing defenses focus on prompt-injection attacks and fail to address threats embedded in tool metadata, leaving MCP-based systems exposed to semantic manipulation. This work analyzes three classes of semantic attacks on MCP-integrated systems: (1) Tool Poisoning, where adversarial instructions are hidden in tool descriptors; (2) Shadowing, where trusted tools are indirectly compromised through contaminated shared context; and (3) Rug Pulls, where descriptors are altered after approval to subvert behavior. To counter these threats, we introduce a layered security framework with three components: RSA-based manifest signing to enforce descriptor integrity, LLM-on-LLM semantic vetting to detect suspicious tool definitions, and lightweight heuristic guardrails that block anomalous tool behavior at runtime. Through evaluation of GPT-4, DeepSeek, and Llama-3.5 across eight prompting strategies, we find that security performance varies widely by model architecture and reasoning method. GPT-4 blocks about 71 percent of unsafe tool calls, balancing latency and safety. DeepSeek shows the highest resilience to Shadowing attacks but with greater latency, while Llama-3.5 is fastest but least robust. Our results show that the proposed framework reduces unsafe tool invocation rates without model fine-tuning or internal modification.

Country of Origin
🇨🇦 Canada

Page Count
32 pages

Category
Computer Science:
Cryptography and Security