KC

kangarooking/cangjie-skill

Developer tools

This page was translated automatically. Technical details may differ from the original. View the original

11.9K stars Quality 74 Trend 89MIT

Distill high-value content from books, long-form videos, podcasts, and more into executable Agent Skills

Overview

A popular idea has recently emerged: distilling colleagues into skills. Even if someone leaves, their experience, tone, and working style will be partially replaced by AI. nuwa-skill is what it does—creating "human skills," like Musk skills and Warren Buffett skills. The matching darwin-skill is responsible for automatically evolving these skills. Distillers are valuable—nuwa-skill has proven this. Distillers** are another dimension: a book, a long interview, a podcast, or a long video on Bilibili or YouTube can all accumulate the methodology the author has spent a long time honing. Rather than imitating someone's way of expressing themselves, breaking down their systematic methodology and turning it into a tool to solve real problems is equally valuable. And there's a real pain point: you might have read a lot of books, saved a lot of videos, or listened to many podcasts, but you just can't put them to use. Especially since major platforms have tons of long, practical videos every day, which are highly time-sensitive and the content is very long; These are often unlikely to have been trained by AI, nor can they be fully absorbed by watching in one go. After distilling this content into skills, AI agents can help you access this knowledge in real scenarios, instead of letting it just sit in notes, favorites, or later gather dust in your list. So cangjie-skill has a clear goal: It's not just for books, but also for videos, podcasts, interviews, speeches, courses, long articles, and datasets with subtitles/transcripts. As long as the content contains extractable, verifiable, and transferable methodologies, cangjie-skill can turn it into an AI skill toolkit that can be called independently, composed, and stress-tested. If you want to distill video content, it's recommended to use it together with the video-downloader skill: first use it to download videos, extract subtitles/audio transcription and key materials, then send the resulting text to cangjie-skill for methodology extraction, skill customization, and stress testing.

Overview (machine translated)

Chinese Simplified · English · Japanese

Cangjie Skill

Distill methodologies from books, long-form videos, and podcasts into callable AI Skills

Finish reading, watching, or listening—and leave with a methodology you can invoke.

Cangjie Skill in 66 seconds

https://github.com/user-attachments/assets/87a2cf7b-2114-48aa-aecd-2640a667db40

Official Website

🌐 Visit the Cangjie Skill official website

The website provides visual Skill Pack browsing, a beginner-friendly usage guide, Skill detail pages, and a contribution submission entry. This GitHub repository remains the sole source for cangjie-skill code, methodology, and templates; the website provides presentation, navigation, and usage guidance.

What’s New in v2.5.0

  • Capability Bundle as the single source of truth: extraction produces stable capability cards and metadata before any installable output is compiled.
  • Two deterministic delivery modes: compile one router-style Skill (single) or a compact pack with a router plus promoted standalone Skills (pack).
  • A unified local toolchain: `scripts/cangjie.

This page was translated automatically. Technical details may differ from the original.

Recommended Tools

Try a different keyword or remove a filter.

Install

npx skillfish add kangarooking/cangjie-skill
New to MCP? Start with the mcpskill.org guide →