--- license: cc-by-4.0 task_categories: - feature-extraction language: - en --- This is a disarmed malware source code dataset. 6032 specimens, 32 languages. It is human-reviewed. It is partly labeled(File Family Vuln Beh Class Pack FUD Unk). We delete the malicious executables but keep the original VirusTotal reports for each specimen. More details can be found in the paper "MASCOT: Analyzing Malware Evolution Through A Well-Curated Source Code Dataset."