Skill Seekers: Auto-Generate Claude Skills

Skill Seekers automates the creation of Claude Skills. Instead of manually copying documentation into prompts, it scrapes websites, GitHub repos, or PDFs and packages them into a format Claude can use as a specialized knowledge base.

Installation

# Recommended
pip install skill-seekers

# For high-performance PDF/OCR support (optional)
pip install skill-seekers[pdf]

Core Workflows

The tool operates in three stages: Scrape → Enhance → Package

1. Scraping Documentation

Using a Preset

skill-seekers scrape --config react

Custom URL

skill-seekers scrape --url https://react.dev --name my-react-skill

GitHub Repo

skill-seekers scrape --github https://github.com/user/repo

2. Unified Multi-Source (Docs + Code)

The most powerful feature for developers. Scrapes documentation AND analyzes source code to detect if docs are outdated.

skill-seekers unified \
  --name my-project \
  --url https://docs.example.com \
  --github https://github.com/example/repo

3. Scraping PDFs

Convert manuals or textbooks into skills:

skill-seekers scrape --pdf path/to/manual.pdf --name technical-skill

Tips:

• Use --ocr for scanned PDFs
• Use --extract-tables for complex data tables

Enhancing and Packaging

Once data is collected, turn it into a .zip file Claude can use.

Enhance

Uses AI to summarize scraped data into a high-quality SKILL.md entry point:

skill-seekers enhance output/my-project/

Package

Creates the final distribution file:

skill-seekers package output/my-project/

Claude Code MCP Integration

Skill Seekers has a built-in Model Context Protocol (MCP) server. Once set up, just ask Claude in natural language:

“Create a skill for the Tailwind documentation.”

“Generate a config for this GitHub repo and package it.”

How LearnAI Team Could Use This

• Convert frequently referenced AI tool documentation into reusable Claude Skills for lesson development.
• Package internal guides as skills so team members can query them consistently inside Claude Code.
• Use the unified docs-plus-code workflow to check whether training examples drift from current source code.

Real-World Use Cases

• A curriculum designer turns React, Tailwind, or OpenAI docs into a Claude Skill before writing a module.
• A developer packages a project repo and docs so Claude can answer implementation questions with local context.
• A training team converts PDF manuals into searchable skills for workshops and support material.

Quick Reference

Example Workflow

# 1. Scrape React docs and repo
skill-seekers unified \
  --name react-skill \
  --url https://react.dev \
  --github https://github.com/facebook/react

# 2. Enhance with AI summary
skill-seekers enhance output/react-skill/

# 3. Package for distribution
skill-seekers package output/react-skill/

# 4. Install to Claude Code
cp output/react-skill.zip ~/.claude/skills/

Now Claude has expert-level React knowledge on demand.