zhangdw/astra-skills-scripts

---
license: apache-2.0
task_categories:
- text-generation
language:
- en
tags:
- agent-skills
- ai-agents
- tool-use
- scripts
- skill-subset
pretty_name: Astra Skills Scripts Subset
size_categories:
- 10K<n<100K
---

# Astra Skills Scripts Subset

## Skill Count: **28,954**

This dataset is a **scripts-focused subset** of `zhangdw/astra-skills`.

It contains only skills whose skill directory includes a `scripts/` subdirectory, intended for researchers studying:

- script-backed agent workflows
- tool invocation patterns around executable helpers
- skill design practices combining prompt instructions with runnable scripts

## Subset Definition

Source dataset: `zhangdw/astra-skills`

Selection rule: keep a skill directory only if it contains a `scripts/` subdirectory.

## Snapshot Summary

- Extraction date: **2026-04-15**
- Source root scanned: `astra-skills/github/`
- Matched skills with scripts: **28,954**
- Copied subset root: `github/`

## Directory Layout

```
github/
<skill_dir>/
SKILL.md
_meta.json
scripts/
...
```

## Relation to Full Dataset

- Full corpus (all deduplicated skills): `zhangdw/astra-skills`
- This dataset is a strict subset filtered by scripts presence.

## Download & Extract

The archive is split into multiple parts for reliable transfer. After downloading, merge and extract with:

```bash
# Download the dataset
huggingface-cli download zhangdw/astra-skills-scripts --repo-type dataset --local-dir astra-skills-scripts

# Merge split parts and extract
cat astra-skills-scripts/astra-skills-scripts-part-*.tar.gz > astra-skills-scripts.tar.gz
tar xzf astra-skills-scripts.tar.gz
```

This produces a `github/` directory containing all skill directories.

## Authors

- Dawei Zhang (GitHub: zhangdw156)

## Citation

If you use this dataset in research, please cite:

```bibtex
@misc{dawei_zhang_2026,
author = { Dawei Zhang },
title = { astra-skills-scripts (Revision 77d9d29) },
year = 2026,
url = { https://huggingface.co/datasets/zhangdw/astra-skills-scripts },
doi = { 10.57967/hf/8428 },
publisher = { Hugging Face }
}
```

## License

Dataset metadata/packaging uses Apache-2.0. Individual skill contents remain subject to their original repository licenses.