A Generative AI Framework for Cloud Security: Automated Attack Simulation and Threat Detection
Mohamed I. D. Helal, Mervat Abu-Elkheir, Maggie Ezzat Mashaly · 2025
In the rapidly evolving domain of cloud computing and cybersecurity, staying ahead of threats requires proactive and adaptive solutions. This paper introduces an end-to-end framework designed to enhance cloud security through automated attack generation, execution, and defense enhancement. The framework consists of four main components: the Database, the Generator, the Executor, and the Defender. The database comprises extensive data sources, including attack logs and security reports. The Generator utilizes the environment and system specifications to generate both traditional and zero-day attack vectors. The Executor classifies attacks, assigns them to specialized workers for execution, and consolidates the results into a comprehensive report. The Defender then analyzes this report to iteratively enhance the Intrusion Detection System (IDS). This iterative process enhances the system’s resilience and robustness to future, unknown threats. The framework is adaptable, allowing customization to meet the security needs of various organizations making it well-suited for applications such as enterprise cloud security, cybersecurity education, and beyond. Future work will focus on building a simplified version of the system, which will later be expanded into a scalable architecture capable of addressing more intricate attack scenarios.