AgentSkillsCN

codifier

当用户需要将项目文档（网站、GitHub 仓库、PDF）转换为 Claude 技能时，可使用此技能。当收到“从文档创建技能”、“将文档编码”、“将 [框架] 文档转换为技能”，或“让 Claude 能够搜索此文档”等请求时，该技能会自动触发。它能够完成文档抓取、GitHub 分析、PDF 提取，以及多源整合的任务。

SKILL.md
--- frontmatter
name: codifier
description: Use when user needs to convert project documentation (websites, GitHub repos, PDFs) into Claude skills. Triggers on "create a skill from docs", "codify documentation", "convert [framework] docs to skill", or "make this documentation searchable for Claude". Handles documentation scraping, GitHub analysis, PDF extraction, and multi-source unification.
version: 1.0.0
dependencies: python>=3.10, git>=2.30

Documentation Codifier

Overview

Converts documentation websites, GitHub repositories, and PDF files into production-ready Claude skills using Skill_Seekers. Transforms 2-4 hours of manual work into 20-40 minutes of automated processing.

Quick Decision Tree

code
User has documentation to codify?
│
├─→ First time using this skill?
│   └─→ Check references/setup.md for installation
│
├─→ Documentation website only?
│   └─→ See "Website Scraping" below
│
├─→ GitHub repository only?
│   └─→ See "GitHub Analysis" below
│
├─→ PDF files only?
│   └─→ See "PDF Extraction" below
│
└─→ Multiple sources (docs + GitHub + PDF)?
    └─→ See "Unified Multi-Source" below

Installation Check

Before proceeding, verify Skill_Seekers is installed:

bash
# Quick check
command -v skill-seekers || python3 -m skill_seekers --version

If not installed: Direct user to references/setup.md for complete installation instructions with uv/venv detection.

Website Scraping

For preset frameworks (React, Vue, Django, FastAPI, etc.):

bash
cd <project-directory>
skill-seekers scrape --config configs/react.json --enhance-local
skill-seekers package output/react/

For custom documentation:

bash
skill-seekers scrape --interactive
# OR
skill-seekers scrape --name myframework --url https://docs.example.com/

Advanced options: See references/advanced-workflows.md

GitHub Analysis

bash
skill-seekers github --repo facebook/react
skill-seekers package output/react/

With authentication (for higher rate limits):

bash
export GITHUB_TOKEN=ghp_your_token_here
skill-seekers github --repo owner/repo

PDF Extraction

bash
skill-seekers pdf --pdf docs/manual.pdf --name myskill
skill-seekers package output/myskill/

Advanced features (tables, OCR, parallel): See references/advanced-workflows.md

Unified Multi-Source

Combines docs + GitHub + PDF with conflict detection:

bash
skill-seekers unified --config configs/react_unified.json
skill-seekers package output/react/

Custom unified config: See references/advanced-workflows.md (Unified Multi-Source section)

Output & Delivery

All workflows create a .zip file in output/<name>.zip. Direct users to:

  1. •Locate the output file
  2. •Upload to Claude at https://claude.ai/skills

Troubleshooting

Command not found: User needs installation - see references/setup.md

No content extracted: CSS selector issue - see references/troubleshooting.md

Rate limiting: Configuration issue - see references/troubleshooting.md

References

  • •references/setup.md - Complete installation with environment detection (uv/venv/pip)
  • •references/advanced-workflows.md - Large docs, async mode, unified multi-source, custom configs, performance optimization
  • •references/troubleshooting.md - Common issues and solutions

For Skill_Seekers deep reference: After installation, refer to cloned Skill_Seekers repository documentation (README.md, docs/CLAUDE.md, and other docs/ files).

Load references as needed based on user context and questions.