CERIAS 2025 Annual Security Symposium


2026 Symposium Posters

Posters > 2026

VERA: A Dual-Validated, Behaviorally-Verified Repository of 15,000+ Raw Windows Ransomware Binaries


PDF

Primary Investigator:
Feng Li

Project Members
Garvit Agarwal, Nithish Vasanth∗, Seunghyun Cho, Yousef Mohammed Y. Alomayri, Noah D. Pumphrey, Feng Li, Yucheng Xie, Xukai Zou
Abstract
Reliable ransomware detection research is hindered by a scarcity of high-quality, reproducible data. Public repositories often suffer from label noise where generic malware, benign files, or corrupted binaries are misclassified. We introduce Validated Execution-based Ransomware Analysis, a repository of behaviorally verified raw Windows ransomware constructed using a comprehensive Dual-Validation Methodology. We processed 40,552 candidates with VirusTotal consensus and subjected them to dynamic verification in the CAPE Sandbox. This process revealed that 23,581 samples were dynamically silent due to defunct infrastructure or incompatibility with modern Windows environments. We segregated these into a secondary VERA-Latent dataset to preserve the comprehensive pool while cleaning the primary set.