Finding RESTful API vulnerabilities using ChatGPT

• Main Author: Ho, Kenneth Jun Minn
• Other Authors: Liu Yang
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Computer and Information Science; RESTful; Vulnerabilities; ChatGPT
• Online Access: https://hdl.handle.net/10356/175120
• Institution: Nanyang Technological University

Modern software applications heavily rely on RESTful APIs for communication and data exchange. Ensuring the reliability and security of these APIs is paramount for robust software development. This project introduces a fully automated testing framework for RESTful APIs. Leveraging advanced technologies such as ChatGPT-enabled instance and sequence generation, and reinforcement learning-driven instance creation, the framework delves into a new form of API testing. The integration of ChatGPT facilitates context-aware test scenario creation, while reinforcement learning enhances adaptability to varying API structures. The project’s main contribution lies in advancing automated testing methodologies, providing a versatile tool that elevates the quality and reliability of RESTful APIs in diverse application scenarios.