An LLM-empowered Automatic Penetration Testing Tool with 8k+ GitHub stars. PentestGPT is designed to automate penetration testing by leveraging the domain knowledge inherent in Large Language Models. It features a three-module architecture (Reasoning, Generation, and Parsing) that emulates human penetration testing workflows. Key Features: Automated penetration testing guidance Integration with multiple LLM backends Real-world evaluation on CTF challenges 228.6% improvement in task completion over baseline models Recognition: Distinguished Artifact Award at USENIX Security 2024 Active collaboration with industry partners (Huawei, ByteDance, etc.)
Aug 1, 2023
A comprehensive study of prompt injection attacks against LLM-integrated applications.
Jun 9, 2023
A comprehensive empirical study of jailbreaking techniques against ChatGPT through prompt engineering.
May 23, 2023
Evaluating ChatGPT's capabilities for automatic code summarization.
May 22, 2023