A Manually-Curated Dataset of Fixes to Vulnerabilities of Open-Source Software

本数据集名为‘A Manually-Curated Dataset of Fixes to Vulnerabilities of Open-Source Software’，由SAP安全研究团队创建，专注于收集和整理开源软件中的漏洞及其修复提交。数据集包含624个漏洞，涉及205个不同的开源Java项目，共有1282个修复提交。创建过程历时四年，通过持续监控安全公告和代码仓库实现。该数据集主要用于训练自动识别代码仓库中安全相关提交的分类器，旨在简化手动识别新漏洞披露修复的复杂问题，提高软件安全性。

This dataset, titled "A Manually-Curated Dataset of Fixes to Vulnerabilities of Open-Source Software", was developed by the SAP Security Research Team. It focuses on collecting and curating vulnerability records and their corresponding fix commits in open-source software. The dataset contains 624 vulnerabilities spanning 205 distinct open-source Java projects, with a total of 1,282 fix commits. It was developed over a four-year period through continuous monitoring of security advisories and code repositories. This dataset is primarily intended for training classifiers that automatically identify security-related commits in code repositories, aiming to simplify the complex manual task of identifying fixes for newly disclosed vulnerabilities and enhancing software security.