Poster: Fuzzing for Command Injections in Medical Software with Large Language Models

In the evolving landscape of medical software security, the threat of command injection looms large, with potential ramifications including compromised patient data and disrupted healthcare services. Addressing these vulnerabilities is paramount, and fuzz testing, or fuzzing, has proven as an indispensable tool for identifying command injection. By identifying and mitigating threats, fuzzing not only guarantees the integrity and reliability of medical software but also upholds patient trust and regulatory adherence, ensuring the continuity and safety of healthcare services.Recent advancements have seen the integration of Large Language Models (LLMs) into security fuzzing, offering a revolutionary approach to uncovering bugs and vulnerabilities. The application of LLMs in fuzzing transcends traditional language-specific limitations, enabling more comprehensive and effective testing across a spectrum of programming languages and systems. In this paper, we exploits the potential of leveraging LLMs for fuzzing medical software against command injection, aiming to harness these advancements to fortify the cybersecurity defenses of critical healthcare technologies.