27 March 2023 in IEEE Access
Malware variants are generated using various evasion techniques to bypass malware detectors, so it is important to understand what properties make them evade malware detection techniques. To do this, a framework is proposed to effectively generate fully-working, unseen malware samples on Windows portable executable (PE) files with various perturbations such as code obfuscation and benign section addition. Using this framework, we were able to bypass various commercial anti-malware solutions (e.g., BitDefender, AVG, Kaspersky, and Avast) using the generated malware variants, with up to 86% more evasiveness than the original malware samples, and up to 28% more evasive compared with our previously proposed solution FUMVar. Our results are useful in terms of improving malware detection techniques, by analyzing different perturbations and their effectiveness, which leads to a better understanding of how malware variants could be generated that are more evasive and which malware categories they belong to. We found that the most effective perturbation is the code obfuscation using XOR – the malware variants generated by the code obfuscation can evade the detection of 28 anti-malware engines on average. Therefore, our experimental results and observations would be useful to develop anti-malware solutions that would be effective in detecting malware variants that have not been seen previously.
17 April 2026 in SSRN Electronic Journal
The rapid growth of complex malware has intensified the burden on security analysts, prompting demand for scalable solutions to automate malware analysis and threat intelligence (CTI) reporting. We introduce CTIGen, a novel framework that automates CTI report generation by integrating large language models (LLMs) with traditional static and dynamic analysis tools. Replicating expert workflows, CTIGen first performs dynamic behavior extraction, then conducts targeted static inspection guided by those behaviors, and finally synthesizes findings into structured, ATT&CK-aligned reports using a graph-based reasoning module. Our evaluation results show that CTIGen achieves strong recall across key tasks: 0.87 for function identification, 0.95 for behavior extraction, 0.36 for function-to-behavior mapping, and 0.58 for IoC identification. Notably, CTIGen identifies malicious functions and their corresponding behaviors within 8 minutes per sample, substantially reducing manual analysis time. Furthermore, CTIGen identifies 121 additional malicious functions with correct behavioral semantics beyond those documented in human-written reports. In ATT&CK technique identification, CTIGen achieves an 77.23% accuracy, demonstrating reliable alignment with annotated ATT&CK techniques. Interestingly, reports generated by CTIGen are more fluent than human-written ones and are rated more useful in a user study. These results demonstrate the potential of LLM-based multi-stage automation for generating accurate and scalable malware threat intelligence.