Dynamická analýza škodlivého softvéru pomocou sandboxu CAPE

Abstract

This thesis focuses on the design and implementation of a system for performing dynamic analysis and subsequently connecting it with GPT-4 for the classification and commenting of malware samples. The theoretical part addresses the issues of dynamic analysis and large language models. In the practical part, a system for conducting dynamic analysis using the CAPEv2 sandbox will be implemented. After obtaining the results from the analysis, they will be parsed into a suitable form and stored. Subsequently, only the necessary data will be selected from the stored data for creating comments and classification of malware using GPT-4.