MALET & Katalina: A Large-Scale Dataset and Static Analysis of macOS Malware

MALET provides a large-scale collection of 89,255 macOS Mach-O binaries, comprising 44,804 malicious samples and 44,457 undetected samples, with 25,835 classified as high-confidence benign via heuristics. The dataset comprises features extracted using the Katalina static analysis framework to extract platform-specific features such as security entitlements, code-signing metadata, and embedded script indicators. Researchers can leverage standardized AVClass2 family and class labels to analyze malware campaigns and identify detection blind spots across the macOS ecosystem. These resources collectively establish a reproducible foundation for advancing macOS malware detection and defense research.

The datasets span 2009-05-22T14:42:53 to 2025-11-22T18:16:49.

MALET是一个包含89255个macOS平台Mach-O二进制文件的大规模数据集，其中包含44804个恶意样本、44457个未检出样本，另有25835个通过启发式方法被归类为高可信度良性样本。该数据集包含通过Katalina静态分析框架提取的各类特征，用于提取平台专属特性，包括安全权限、代码签名元数据以及嵌入式脚本标识。研究人员可借助标准化的AVClass2家族及类别标签，开展恶意软件活动分析，并识别macOS生态系统中的检测盲区。上述资源共同为推进macOS恶意软件检测与防御研究构建了可复现的研究基础。

该数据集的时间跨度为2009年5月22日14时42分53秒至2025年11月22日18时16分49秒。