skill-diffs
What it is
A large-scale dataset of agent skill (SKILL.md) commit histories and diffs scraped from public GitHub repositories across four platforms. It provides structured (before, after, intent) pairs and quality metrics specifically for fine-tuning skill-editing or curator models.
The problem it solves
Lack of high-quality, evolution-focused training data for agent skills, as existing datasets typically only capture static snapshots rather than the iterative refinement process.
What you could use it for
- Fine-tuning LLMs to act as skill editors or curators
- Training DPO or preference models on skill evolution
- Benchmarking models on skill patching tasks
- Mining patterns in how agent skills are iteratively refined
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.