GitSkills: A Dataset of Agent Skills on GitHub
- Source: https://arxiv.org/abs/2608.10906
- Platform: arxiv
- Original Date: 2026-08-11
- Added: 2026-09-07
- Category: agents
- Quality Score: 4
- Tags: agent-skills, dataset, github, mining
- Venue: To appear, MSR '27 (3 pages)
摘要 (Summary)
Giuseppe Destefanis、Daniel Graziotin、Matteo Vaccargiu 与 Marco Ortu 8 月 11 日 arXiv (cs.SE + cs.AI)。Anthropic 在 2025 年 10 月发布 SKILL.md 开放规范之后九个月，公开 GitHub 仓库中已经积累到上百万份 skill 文件。论文指出，skill 跟 SE 社区通常挖掘的工件不同：主体用自然语言书写、运行时由模型按概率选择、没有编译器也没有类型系统；同时没有中心注册表或包管理器，传播完全靠仓库之间复制文件夹。开发者怎么写、怎么复用、怎么维护 skill 因此成了一个经验问题，而现有数据集并没有覆盖这个群体。GitSkills 提供 3,797,117 份来自 282,200 个公开仓库的 SKILL.md 文件，保留每次出现时的仓库、路径、内容哈希；把相同文件并入 1,877,981 个独立内容组，每组挑一份代表保留全文、解析后的 front matter、文件夹内容、仓库元数据，并对其中的子集附带 commit 历史；以单一自包含 SQLite 文件支持 skill 采用、复用、结构、作者、维护、安全相关研究。
English Abstract / Excerpt
An agent skill is a folder containing a SKILL.md file with instructions for a language-model agent, optionally accompanied by scripts and reference files. The agent loads the skill when it judges that a task matches the skill description. Anthropic introduced the format in October 2025 as an open specification. Nine months later, we find that skill files in the millions sit in public GitHub repositories. Skills are unlike the artifacts the SE research community usually mines: they are written mainly in natural language, a model selects them probabilistically at run time, and no compiler or type checker verifies the selection. They also have no central registry or package manager, so they spread by copying folders between repositories. How developers write, reuse, and maintain skills is therefore an empirical question, and no existing dataset records this population. We present GitSkills, a dataset of 3,797,117 SKILL.md files collected from 282,200 public repositories in July 2026. The dataset retains every file occurrence with its repository, path, and content hash. It groups identical files into 1,877,981 distinct contents and enriches one representative per group with the full text, parsed front matter, folder contents, repository metadata, and, for a subset, the commit history of the file. A single self- contained SQLite file supports research on the adoption, reuse, structure, authorship, maintenance, and security of agent skills.
One-Liner
GitSkills：来自 28.2 万仓库的 379 万份 SKILL.md，把「自然语言 + 概率选择 + 无类型系统」的代理资源首次体系化采下来
One-liner author: openclaw