LLM-Based Identification of Infostealer Infection Vectors from Screenshots: The Case of Aurora

Infostealers exfiltrate credentials, session cookies, and sensitive data from infected systems. With over 29 million stealer logs reported in 2024, manual analysis and mitigation at scale are virtually unfeasible/unpractical. While most research focuses on proactive malware detection, a significant gap remains in leveraging reactive analysis of stealer logs and their associated artifacts. Specifically, infection artifacts such as screenshots, image captured at the point of compromise, are largely overlooked by the current literature. This paper introduces a novel approach leveraging Large Language Models (LLMs), more specifically gpt-4o-mini, to analyze infection screenshots to extract potential Indicators of Compromise (IoCs), map infection vectors, and track campaigns. Focusing on the Aurora infostealer, we demonstrate how LLMs can process screenshots to identify infection vectors, such as malicious URLs, installer files, and exploited software themes. Our method extracted 337 actionable URLs and 246 relevant files from 1000 screenshots, revealing key malware distribution methods and social engineering tactics. By correlating extracted filenames, URLs, and infection themes, we identified three distinct malware campaigns, demonstrating the potential of LLM-driven analysis for uncovering infection workflows and enhancing threat intelligence. By shifting malware analysis from traditional log-based detection methods to a reactive, artifact-driven approach that leverages infection screenshots, this research presents a scalable method for identifying infection vectors and enabling early intervention.

The Use of Large Language Models (LLM) for Cyber Threat Intelligence (CTI) in Cybercrime Forums

Large language models (LLMs) can be used to analyze cyber threat intelligence (CTI) data from cybercrime forums, which contain extensive information and key discussions about emerging cyber threats. However, to date, the level of accuracy and efficiency of LLMs for such critical tasks has yet to be thoroughly evaluated. Hence, this study assesses the accuracy of an LLM system built on the OpenAI GPT-3.5-turbo model [7] to extract CTI information. To do so, a random sample of 500 daily conversations from three cybercrime forums, XSS, Exploit.in, and RAMP, was extracted, and the LLM system was instructed to summarize the conversations and code 10 key CTI variables, such as whether a large organization and/or a critical infrastructure is being targeted. Then, two coders reviewed each conversation and evaluated whether the information extracted by the LLM was accurate. The LLM system performed strikingly well, with an average accuracy score of 98%. Various ways to enhance the model were uncovered, such as the need to help the LLM distinguish between stories and past events, as well as being careful with verb tenses in prompts. Nevertheless, the results of this study highlight the efficiency and relevance of using LLMs for cyber threat intelligence.