Language Model-Driven Agent-Based Framework for Generating and Evaluating Cyber Attack Scenarios

As cybersecurity threats have become increasingly complex, conventional rule-based detection systems struggle to keep pace. Existing cybersecurity training simulators rely on manually scripted scenarios and rigid operational flows, resulting in limited scalability and low robustness against novel attack strategies. To overcome these challenges, this study proposes a cyber-attack scenario generation framework based on an agent-based system. The framework is composed of two attack agents that construct threat scenarios and an evaluation agent that analyzes the effectiveness of the generated scenarios. Our fine-tuned language model agent generates technically coherent scenarios rooted in domain knowledge, whereas our prompt-engineered large language model attack agent produces flexible and diverse scenarios through structured reasoning. The evaluation agent supports objective and multi-perspective validation, ensuring both consistency and feasibility. The agent-based framework offers a scalable foundation for generating practical threat simulations that are aligned with real-world complexities.