QUT-DV25
收藏资源简介:
QUT-DV25是一个用于动态分析下一代软件供应链攻击的数据集,由昆士兰科技大学的研究团队创建。该数据集包含14271个Python包的行为轨迹,其中7127个表现出恶意行为。这些包在隔离的沙盒环境中执行,使用扩展的Berkeley Packet Filter (eBPF)内核和用户级探针进行实时监控,记录了36个实时特征,包括系统调用、网络流量、资源使用、目录访问模式、依赖日志和安装行为。这些特征使得研究下一代攻击向量成为可能,如多阶段恶意软件执行、远程访问激活和动态有效载荷生成。QUT-DV25数据集在恶意检测系统中表现出色,为在快速发展的软件供应链生态系统中开发和基准化高级威胁检测提供了坚实的基础。
QUT-DV25 is a dataset for dynamic analysis of next-generation software supply chain attacks, created by a research team at Queensland University of Technology. The dataset contains behavioral traces of 14,271 Python packages, among which 7,127 exhibit malicious behaviors. These packages were executed in an isolated sandbox environment, with real-time monitoring conducted using extended Berkeley Packet Filter (eBPF) kernel and user-level probes, capturing 36 real-time features including system calls, network traffic, resource usage, directory access patterns, dependency logs, and installation behaviors. These features enable research into next-generation attack vectors such as multi-stage malware execution, remote access activation, and dynamic payload generation. The QUT-DV25 dataset delivers outstanding performance in malicious detection systems, providing a solid foundation for developing and benchmarking advanced threat detection within the rapidly evolving software supply chain ecosystem.




