Sean Oesch

As large language models are increasingly deployed in sensitive environments, fingerprinting attacks pose significant privacy and security risks. We present a study of LLM fingerprinting from both offensive and defensive perspectives. Our attack methodology uses reinforcement learning to automatically optimize query selection, achieving better fingerprinting accuracy with only 3 queries compared to randomly selecting 3 queries from the same pool. Our defensive approach employs semantic-preserving output filtering through a secondary LLM to obfuscate model identity while maintaining semantic integrity. The defensive method reduces fingerprinting accuracy across tested models while preserving output quality. These contributions show the potential to improve fingerprinting tools capabilities while providing practical mitigation strategies against fingerprinting attacks.

Modern security operation centers (SOCs) rely on operators and a tapestry of logging and alerting tools with large scale collection and query abilities. SOC investigations are tedious as they rely on manual efforts to query diverse data sources, overlay related logs, and correlate the data into information and then document results in a ticketing system. Security orchestration, automation, and response (SOAR) tools are a new technology that promise to collect, filter, and display needed data; automate common tasks that require SOC analysts' time; facilitate SOC collaboration; and, improve both efficiency and consistency of SOCs. SOAR tools have never been tested in practice to evaluate their effect and understand them in use. In this paper, we design and administer the first hands-on user study of SOAR tools, involving 24 participants and 6 commercial SOAR tools. Our contributions include the experimental design, itemizing six characteristics of SOAR tools and a methodology for testing them. We describe configuration of the test environment in a cyber range, including network, user, and threat emulation; a full SOC tool suite; and creation of artifacts allowing multiple representative investigation scenarios to permit testing. We present the first research results on SOAR tools. We found that SOAR configuration is critical, as it involves creative design for data display and automation. We found that SOAR tools increased efficiency and reduced context switching during investigations, although ticket accuracy and completeness (indicating investigation quality) decreased with SOAR use. Our findings indicated that user preferences are slightly negatively correlated with their performance with the tool; overautomation was a concern of senior analysts, and SOAR tools that balanced automation with assisting a user to make decisions were preferred.

In this paper, we present a scientific evaluation of four prominent malware detection tools to assist an organization with two primary questions: To what extent do ML-based tools accurately classify previously- and never-before-seen files? Is it worth purchasing a network-level malware detector? To identify weaknesses, we tested each tool against 3,536 total files (2,554 or 72\% malicious, 982 or 28\% benign) of a variety of file types, including hundreds of malicious zero-days, polyglots, and APT-style files, delivered on multiple protocols. We present statistical results on detection time and accuracy, consider complementary analysis (using multiple tools together), and provide two novel applications of the recent cost-benefit evaluation procedure of Iannacone \& Bridges. While the ML-based tools are more effective at detecting zero-day files and executables, the signature-based tool may still be an overall better option. Both network-based tools provide substantial (simulated) savings when paired with either host tool, yet both show poor detection rates on protocols other than HTTP or SMTP. Our results show that all four tools have near-perfect precision but alarmingly low recall, especially on file types other than executables and office files -- 37% of malware tested, including all polyglot files, were undetected. Priorities for researchers and takeaways for end users are given.

Standardized file formats play a key role in the development and use of computer software. However, it is possible to abuse standardized file formats by creating a file that is valid in multiple file formats. The resulting polyglot (many languages) file can confound file format identification, allowing elements of the file to evade analysis.This is especially problematic for malware detection systems that rely on file format identification for feature extraction. File format identification processes that depend on file signatures can be easily evaded thanks to flexibility in the format specifications of certain file formats. Although work has been done to identify file formats using more comprehensive methods than file signatures, accurate identification of polyglot files remains an open problem. Since malware detection systems routinely perform file format-specific feature extraction, polyglot files need to be filtered out prior to ingestion by these systems. Otherwise, malicious content could pass through undetected. To address the problem of polyglot detection we assembled a data set using the mitra tool. We then evaluated the performance of the most commonly used file identification tool, file. Finally, we demonstrated the accuracy, precision, recall and F1 score of a range of machine and deep learning models. Malconv2 and Catboost demonstrated the highest recall on our data set with 95.16% and 95.45%, respectively. These models can be incorporated into a malware detector's file processing pipeline to filter out potentially malicious polyglots before file format-dependent feature extraction takes place.