Breaking MCP with Function Hijacking Attacks: Novel Threats for Function Calling and Agentic Models

TL;DR

This paper introduces a novel function hijacking attack (FHA) on agentic LLMs that manipulates function invocation, demonstrating its robustness and broad applicability across models and domains, highlighting security vulnerabilities.

Contribution

The paper presents FHA, a new attack method that is model-agnostic and capable of hijacking function calls in agentic LLMs, revealing significant security concerns.

Findings

FHA achieves 70% to 100% attack success rate across 5 models.

FHA is effective regardless of context semantics or function set.

Experiments highlight the need for improved security measures in agentic systems.

Abstract

The growth of agentic AI has drawn significant attention to function calling Large Language Models (LLMs), which are designed to extend the capabilities of AI-powered system by invoking external functions. Injection and jailbreaking attacks have been extensively explored to showcase the vulnerabilities of LLMs to user prompt manipulation. The expanded capabilities of agentic models introduce further vulnerabilities via their function calling interface. Recent work in LLM security showed that function calling can be abused, leading to data tampering and theft, causing disruptive behavior such as endless loops, or causing LLMs to produce harmful content in the style of jailbreaking attacks. This paper introduces a novel function hijacking attack (FHA) that manipulates the tool selection process of agentic models to force the invocation of a specific, attacker-chosen function. While existing attacks focus on semantic preference of the model for function-calling tasks, we show that FHA is largely agnostic to the context semantics and robust to the function sets, making it applicable across diverse domains. We further demonstrate that FHA can be trained to produce universal adversarial functions, enabling a single attacked function to hijack tool selection across multiple queries and payload configurations. We conducted experiments on 5 different models, including instructed and reasoning variants, reaching 70% to 100% ASR over the established BFCL dataset. Our findings further demonstrate the need for strong guardrails and security modules for agentic systems.