How to Set Up the kangarooking/cangjie-skill Project Locally

You can set up the kangarooking/cangjie-skill project locally by cloning the repository and ensuring Python 3.9 or higher is installed, requiring no external dependencies or compiled binaries.

The kangarooking/cangjie-skill repository is a Python-centric, template-driven pipeline that transforms books, videos, and podcasts into agent-callable AI skills using the RIA-TV++ methodology. Setting up this project locally requires only a modern Python interpreter and the standard library, making it lightweight and accessible for immediate use. This guide walks you through the complete local setup process, from cloning the repository to validating your environment with the built-in tooling.

Prerequisites

Before you begin, ensure you have the following:

• Python 3.9 or higher installed and available on your $PATH
• (Optional) A GitHub personal access token – only required if you want to run the star history generation script located at scripts/generate_star_history.py

The repository contains no compiled binaries and relies exclusively on the Python standard library, so no requirements.txt or virtual environment setup is necessary.

Step-by-Step Local Setup

Clone the Repository

Start by cloning the repository to your local machine:

git clone https://github.com/kangarooking/cangjie-skill.git
cd cangjie-skill

Verify you have the expected directory structure by listing the contents:

ls -R

You should see directories including extractors/, methodology/, scripts/, templates/, and assets/, along with the README.md and SKILL.md files at the root.

Configure Optional GitHub Access

If you plan to run the star history generator, export your GitHub token as an environment variable:

export GITHUB_TOKEN=ghp_XXXXXXXXXXXXXXXXXXXXXXXXXXXX

Never commit this token to version control. This token is only required for the generate_star_history.py script and is not needed for the core skill generation pipeline.

Validate Your Setup with the Star History Script

Run the star history generator to validate your Python environment and demonstrate the repository's built-in tooling:

python3 scripts/generate_star_history.py \
    --repo kangarooking/cangjie-skill \
    --output assets/star-history.svg

This script uses only the Python standard library (urllib, concurrent.futures, and json) to fetch repository statistics. It reads base64-encoded assets from assets/xkcd.woff.b64 and assets/star-history-logo.png.b64 to generate a self-contained SVG visualization. If the script executes successfully and creates assets/star-history.svg, your local setup is fully functional.

Understanding the Project Architecture

Once set up, familiarize yourself with the three-layer architecture defined in the source code:

1. Content Ingestion & Extraction – Located in extractors/, this layer contains five parallel extractors defined in framework-extractor.md, principle-extractor.md, case-extractor.md, counter-example-extractor.md, and glossary-extractor.md. These files define prompts that extract structured data from raw content.

2. RIA-TV++ Processing – The methodology/ directory contains the seven-stage pipeline overview in 00-overview.md and subsequent stage files (01-stage0-adler.md through 07-stage5-deliver.md). These define the Reading, Interpretation, and Appropriation methodology combined with Triple Verification and Execution.

3. Delivery & Integration – The templates/ directory contains Jinja-style templates including SKILL.md.template, INDEX.md.template, DIGEST.md.template, and BOOK_OVERVIEW.md.template that generate the final skill artifacts.

Generate a Skill from Template

To test the template system locally, use the following Python snippet to render a skill markdown file:

from pathlib import Path
import jinja2

# Load the SKILL template

template_path = Path("templates/SKILL.md.template")
template = jinja2.Environment(
    loader=jinja2.FileSystemLoader(template_path.parent)
).get_template(template_path.name)

# Render a minimal example (replace placeholders with real values)

rendered = template.render(
    **{
        "skill-slug": "example-skill",
        "BOOK_TITLE": "Example Book",
        "AUTHOR": "Jane Doe",
        "章节": "Chapter 1",
        "tag1": "framework",
        "tag2": "principle",
        "Skill Title": "Example Skill",
        "原文引用": "“A concise quote from the source.”",
        "CHAPTER": "1",
        "案例名": "Case Study",
        "作者遇到了什么": "Problem description",
        "作者怎么用这个方法论思考": "Method application",
        "得出了什么": "Conclusion",
        "实际发生了什么": "Result",
        "场景 1 — 具体到可识别的情况": "User is deciding on X",
        "典型措辞 1": "How do I X?",
        "related-skill-a": "another-skill",
        "反场景 1 — 为什么不适用": "When Y is true",
        "DATE": "2026-07-17",
    }
)

print(rendered)

This example demonstrates how templates/SKILL.md.template can be programmatically filled to produce a ready-to-install skill file according to the cangjie-skill execution specification.

Install Skills into Claude Code

After generating skill files, install them manually into your AI agent environment:

1. Locate the generated *.md file (e.g., example-skill/SKILL.md)
2. Copy the entire skill directory into Claude Code's skills/ folder
3. Restart Claude Code – the skill will appear in the UI and can be invoked via the specified triggers

Summary

• Clone the repository from https://github.com/kangarooking/cangjie-skill.git to access the pipeline
• Verify Python 3.9+ is installed; no external packages are required
• Test your setup by running python3 scripts/generate_star_history.py with an optional GITHUB_TOKEN
• Explore the architecture through methodology/00-overview.md and the extractors/ directory
• Generate skills using Jinja2 templates found in templates/SKILL.md.template
• Install completed skills into Claude Code by copying them to the skills/ directory

Frequently Asked Questions

What Python version is required for cangjie-skill?

The project requires Python 3.9 or higher. According to the source code analysis, the repository uses only the Python standard library, so any modern Python 3.9+ interpreter will work without additional package installation.

Do I need to install external dependencies or create a virtual environment?

No. The kangarooking/cangjie-skill project contains no compiled binaries and requires no external packages. All functionality relies on the Python standard library, making virtual environments optional rather than mandatory for local setup.

How do I use the star history generator script?

Export a GITHUB_TOKEN environment variable, then execute python3 scripts/generate_star_history.py --repo kangarooking/cangjie-skill --output assets/star-history.svg. The script uses urllib and concurrent.futures from the standard library to fetch repository data and generates an SVG using base64-encoded assets from the assets/ directory.

Where are the skill templates located?

Skill templates are located in the templates/ directory at the repository root. The primary template is SKILL.md.template, which uses Jinja2 syntax for variable substitution. Additional templates include INDEX.md.template, DIGEST.md.template, and BOOK_OVERVIEW.md.template for different output formats.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}