Back to skills
SKILL.md
Skills Builder
ASecurityBuild Claude Code skills from documentation websites. Scrapes Mintlify, ReadTheDocs/Sphinx, Docusaurus, GitBook, and static HTML sites. Generates properly formatted skills with SKILL.md and reference files. Use when: creating skills from docs, scraping documentation, generating reference materials from URLs.
- 207 stars
- 0 votes
- 0 copies
- 3 views
- Added February 7, 2026
Works with
Security analysis
96/100- Installs packages at runtime which could introduce malicious dependencies
Pro scans all 9 files and shows the line behind each finding
npx -y skills add NeverSight/skills_feed --skill skills-builder --agent claude-codeAre you the author of Skills Builder?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/neversight-skills-builder)---
name: "skills-builder"
description: "Build Claude Code skills from documentation websites. Scrapes Mintlify, ReadTheDocs/Sphinx, Docusaurus, GitBook, and static HTML sites. Generates properly formatted skills with SKILL.md and reference files. Use when: creating skills from docs, scraping documentation, generating reference materials from URLs."
---
# Skills Scraper
Build Claude Code skills directly from documentation websites.
## Quick Start
```bash
pip install requests beautifulsoup4 lxml playwright
playwright install chromium
python scripts/build.py https://mintlify.com/docs mintlify-docs
```
## Supported Platforms
| Platform | Description | Browser |
| ------------- | ------------------------ | ------- |
| `mintlify` | Mintlify-powered docs | Yes |
| `readthedocs` | ReadTheDocs/Sphinx sites | No |
| `docusaurus` | Docusaurus/React docs | Yes |
| `gitbook` | GitBook-powered sites | Yes |
| `generic` | Static HTML sites | No |
| `auto` | Auto-detect (default) | - |
## Usage
```bash
python scripts/build.py <url> <name> [options]
```
**Options:**
- `-o, --output` - Output directory (default: ./skills)
- `-p, --platform` - Platform (default: auto)
- `--max-pages N` - Limit pages (0 = unlimited)
- `--max-depth N` - Max crawl depth (default: 10)
- `--no-browser` - Disable Playwright
- `--no-sitemap` - Disable sitemap discovery
- `-v, --verbose` - Verbose output
## Examples
```bash
# Mintlify
python scripts/build.py https://resend.com/docs resend-api
# ReadTheDocs/Sphinx
python scripts/build.py https://flask.palletsprojects.com/en/stable/ flask-docs
# Docusaurus
python scripts/build.py https://reactnative.dev/docs/getting-started react-native
# GitBook
python scripts/build.py https://docs.cortex.io cortex-docs
# Static HTML
python scripts/build.py https://docs.python.org/3/ python-docs -p generic
```
## Output Structure
```
skill-name/
├── SKILL.md # Metadata + instructions
└── references/ # Documentation files
├── api.md
├── getting-started.md
└── ...
```
## Requirements
- Python 3.10+
- requests, beautifulsoup4, lxml
- playwright (for JS-rendered sites)
Files in this skill
- SKILL.md
- description_cn.txt
- description_de.txt
- description_en.txt
- description_es.txt
- description_fr.txt
- description_ja.txt
- description_ko.txt
- description_tw.txt
Attribution
Comments
Loading comments…