Cangjie Skill — distill methodologies from books, videos & podcasts into callable AI Skills
Distill methodologies from books, long-form videos, podcasts, courses, and document collections into independently callable, composable, and pressure-tested AI skill packs. Uses the RIA-TV++ pipeline to transform source texts into a reusable Capability Bundle, then compiles that source into installable skills. Works with OpenClaw, Claude Code, and DeepSeek Harness.
Prompt template
$cangjie-distill Distill this book into a set of executable Agent Skills: <file path>
How to use
- 1Prepare source text: a book, video transcript, podcast transcript, interview notes, course, or document collection
- 2Run the RIA-TV++ pipeline — start with Whole-Content Comprehension (Adler Analysis) producing BOOK_OVERVIEW.md
- 3Run five parallel extractors (frameworks, principles, cases, counter-examples, glossary) to pull candidate units
- 4Apply Triple Verification + Promotion Gate — each candidate passes evidence checks before earning an independent entrypoint
- 5Build RIA++ capability cards (R/I/A1/A2/E/B) inside .cangjie/capabilities/
- 6Encode dependencies, contrasts, and compositions via Zettelkasten linking
- 7Pressure-test each skill with bait questions and cross-skill confusion tests
- 8Compile deterministically to single or pack delivery mode, alongside a reader-facing DIGEST.md and installable artifacts
Pro tips
- Works with books, videos with subtitles/transcripts, podcasts, interviews, talks, courses, long-form articles, and document collections
- Pair with the video-downloader skill to download videos, extract subtitles/transcripts, then feed the text to cangjie-skill
- Not every content source deserves a standalone skill — strict filtering is built into the pipeline
- Generated packs include BOOK_OVERVIEW.md, INDEX.md, DIGEST.md, GLOSSARY.md, multiple SKILL.md files, and test-prompts.json
- Proven outputs include Buffett Letters (20 skills), Poor Charlie's Almanack (12 skills), and AI for Everyone (25 skills)
- Available as a standalone plugin for DeepSeek Harness (dsh)
Works best with
Ratings & Reviews
Loading reviews…
Related Skills
Prompt Chaining — break complex tasks into sequential steps
Decompose a complex task into a sequence of simpler prompts, where each step's output feeds into the next step's input.
Avoid AI Writing — Audit & rewrite content to remove AI writing patterns
A practical skill for any AI agent to audit and rewrite content and remove AI writing patterns ('AI-isms'). Three modes: Rewrite (default, flags and rewrites with built-in second pass), Detect (flags without rewriting, shows real problems vs judgment calls), and Edit (edits prose files in place with minimal targeted changes). 112-entry word replacement table across 3 tiers + 10 Tier 3 phrases, 69 pattern categories. Optional voice profiles (casual/professional/technical/warm/blunt). Works with Claude Code, OpenClaw, Hermes, Cursor, Codex, and any agentskills.io-compatible agent.
Codebase to Course — Turn any GitHub repo into an interactive HTML course
An Agent Skill that turns any GitHub codebase into a beautiful, interactive HTML course. Give it a repo URL and it produces a self-contained course website with lessons, code walkthroughs, quizzes, and progress tracking. Perfect for onboarding new developers, creating documentation, or learning from open source projects. The course is generated as static HTML that can be hosted anywhere — GitHub Pages, Netlify, Vercel, or locally. No build step, no dependencies, just open index.html.