Skip to content
Back to search

cangjie-skill

SkillWorks with Claude CodeGitHub

Distill a book, long-video transcript, podcast, course, or interview into a coherent set of executable skills. Use when the user asks to "拆书" / "蒸馏一本书" / "把 XX 书做成 skill" / "把这个视频/播客/课程蒸馏成 skill" / "turn a book or video into skills" — i.e. wants the frameworks, principles, and methodologies in long-form content extracted into atomic, reusable Claude skills that an agent can invoke in real-world situations. NOT for simple summarization, book reviews, or role-playing as the author (that is nuwa-skill's job).

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/kangarooking-cangjie-skill/badge.svg)](/listings/kangarooking-cangjie-skill)

HTML

<a href="/listings/kangarooking-cangjie-skill"><img src="/api/listings/kangarooking-cangjie-skill/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

SKILL.md · Manifest · 184 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. ---
  2. name: cangjie-skill
  3. description: Distill a book, long-video transcript, podcast, course, or interview into a coherent set of executable skills. Use when the user asks to "拆书" / "蒸馏一本书" / "把 XX 书做成 skill" / "把这个视频/播客/课程蒸馏成 skill" / "turn a book or video into skills" — i.e. wants the frameworks, principles, and methodologies in long-form content extracted into atomic, reusable Claude skills that an agent can invoke in real-world situations. NOT for simple summarization, book reviews, or role-playing as the author (that is nuwa-skill's job).
  4. metadata:
  5. cangjie.version: "2.5.0"
  6. ---
  7. # cangjie-skill — 把一本书蒸馏成一组可执行 skills 的元 skill
  8. ## 使命
  9. 把一本书里沉淀的方法论,拆解成**原子化、可被 agent 在真实场景下调用**的能力,并按用户目的编译成合适数量的 skill,让读者真正用起来。
  10. > **术语约定**: 本文档及 `methodology/`、`extractors/` 中所有的"书",泛指一切被蒸馏的长内容 — 书籍、长视频转写、播客文字稿、课程、访谈、长文、资料集。
  11. **边界**:
  12. - ✅ 做: 方法论 / 决策框架 / 操作流程 / 计算规则 / 排障 / 清单 / 原则的蒸馏,概念体系作为参考支撑
  13. - ❌ 不做: 书摘 / 读后感 / 作者人设角色扮演 (后者请用 nuwa-skill)
  14. ## 核心方法论: RIA-TV++(v2.5 Bundle 版)
  15. 一个五阶段 + 并行提取 + 三重验证 + 晋级门 + darwin 兼容测试的流水线。详见 `methodology/00-overview.md`。
  16. ```
  17. 阶段 0: Adler 整书理解 → BOOK_OVERVIEW.md
  18. 阶段 1: 5 个 agent 并行提取 → 候选方法论单元池
  19. 阶段 1.5: 三重验证筛选(知识验证) → 通过的单元 (用户轻确认)
  20. 阶段 1.6: 独立 Skill 晋级门(产品化验证)→ promoted / router 去向
  21. 阶段 2: RIA++ 构造能力卡 → .cangjie/capabilities/cards/<slug>.md
  22. 阶段 3: Zettelkasten 链接 → verified.yaml 的 also_read + GLOSSARY
  23. 阶段 4: 压力测试 (darwin 兼容) → 评测用例 + 回炉淘汰
  24. 阶段 5: 编译与交付 → cangjie.py compile(single/pack)+ DIGEST.md + 安装
  25. ```
  26. **v2.5 关键变化(ADR-002)**: 阶段 2–4 不再直接把每个单元写成独立的最终 Skill 目录,而是产出一份 **Capability Bundle**(`books/<slug>/.cangjie/capabilities/verified.yaml` + `cards/*.md`)。single 与 pack 都从这同一份 Bundle 由 `scripts/cangjie.py compile` 确定性编译,两者引用同一组稳定 `capability_id`。旧的 one-to-one 输出仍受支持(legacy-pack,见 `docs/migrations/2026-08-25-v2.0-to-v2.1.md`)。
  27. ## 何时调用此 skill
  28. 用户说类似:
  29. - "帮我拆《穷查理宝典》"
  30. - "把毛选蒸馏成 skill"
  31. - "把这个 B 站视频/播客/课程蒸馏成 skill"
  32. - "distill this book into skills: <path>"
  33. - "我想把这本书的方法论做成可用的 skill"
  34. ## 输入要求
  35. 在开始前**必须**从用户处确认:
  36. > **脚本依赖**: 确定性脚本(`scripts/`)需要 Python 3.10+ 与 PyYAML(`python3 -m pip install pyyaml`)。缺 PyYAML 时先运行 `python3 scripts/cangjie.py doctor` 自检,它会给出安装指引。
  37. 1. **内容文本来源**: PDF / EPUB / TXT / 字幕文件 / 转写稿路径, 或可访问的纯文本。**不要**在没有文本的情况下"凭记忆"蒸馏 — 宁可停下来问用户要。(视频/播客建议先用 video-downloader 类工具拿到转写文本)
  38. 2. **内容元信息**: 书籍是"书名 + 作者 + 出版年"; 视频/播客/课程是"标题 + 作者(UP 主/主播/讲者) + 发布时间"。用于目录命名和审计。
  39. 3. **使用目的**(决定输出模式推荐): 学习/查阅这本书 → 倾向 single;接入日常工作流、跨书组合 → 倾向 pack。不确定时按 single-first 原则先推荐 single。
  40. 4. **是否首次试点**: 如果用户是第一次用 cangjie-skill,建议先蒸馏 1 份内容验证流程再批量。
  41. **非书籍内容的字段映射**: 章节类字段对视频填时间戳或分 P,对播客填集数,对课程填讲次 — 保证可追溯即可。
  42. ## 输出结构
  43. ```
  44. books/<book-slug>/
  45. ├── PIPELINE_STATE.md # 流水线状态: 当前阶段 + 进度 (断点续跑用)
  46. ├── BOOK_OVERVIEW.md # 阶段 0 产出: 主旨/骨架/术语/批判
  47. ├── verified.md # 阶段 1.5 产出: 通过三重验证的单元 + 判定理由
  48. ├── coverage-audit.md # 原书关键任务 → 候选 → 判定 → 实际交付去向
  49. ├── references.md # 有依据但不执行的参考内容及交付映射
  50. ├── needs-review.md # 来源/条件/测试缺口;不编译为 active 能力
  51. ├── GLOSSARY.md # 阶段 3 产出: 全书共享术语词典
  52. ├── DIGEST.md # 阶段 5 产出: 面向读者的精华长文
  53. ├── candidates/ # 阶段 1 产出: 原始候选池 (审计用)
  54. ├── rejected/ # 阶段 1.5 淘汰的单元 + 原因 (审计用)
  55. └── .cangjie/ # v2.5 侧车层(编译事实源 + 运行记录)
  56. ├── capabilities/
  57. │ ├── verified.yaml # Capability Bundle(唯一编译事实源, capability-bundle.schema.json)
  58. │ ├── cards/<slug>.md # RIA 能力卡(R/I/A1/A2/E/B, 阶段 2 产出)
  59. │ ├── resources/ # 能力显式声明的 UTF-8 脚本、CSV/JSON/Markdown 模板(可选)
  60. │ ├── destinations.json # 晋级/路由去向映射(阶段 1.6 产出)
  61. │ └── book/{overview.md,glossary.md}
  62. ├── runs/<run-id>/ # 每次编译/更新的决策报告与校验日志
  63. └── snapshots/ # 发布前快照(rollback 用)
  64. ```
  65. 最终交付物由 `scripts/cangjie.py compile` 从 Bundle 编译(--output single 得到 1 个入口 + 能力卡;--output pack 得到 1 个来源路由入口 + 少量晋级 Skill)。
  66. ## 执行流程 (严格按顺序)
  67. **断点续跑**: 开始前先检查 `books/<slug>/PIPELINE_STATE.md` 是否存在。存在则读取并从记录的阶段续跑,不要从头重来。每完成一个阶段,更新该文件。
  68. ### 阶段 0 — 整书理解
  69. 1. 读取用户提供的书本文本。大文件分块阅读。
  70. 2. 执行 `methodology/01-stage0-adler.md` 中的 Adler 四步 (结构 / 解释 / 批判 / 应用)。
  71. 3. 按 `templates/BOOK_OVERVIEW.md.template` 填充,写入 `books/<slug>/BOOK_OVERVIEW.md`。
  72. 4. 把产出展示给用户确认:"骨架我理解对了吗?有没有你希望重点突出的方向?" 得到确认再进入阶段 1。
  73. ### 阶段 1 — 5 个 sub-agent 并行提取
  74. **并行** spawn 5 个 Task sub-agents(使用 Agent 工具,一次调用中发起 5 个):
  75. | sub-agent | 读取的 prompt | 上下文策略 | 产出 |
  76. |---|---|---|---|
  77. | 框架提取器 | `extractors/framework-extractor.md` | **全量扫描**(兼顾跨章节与单处完整机制) | 决策框架 / 思维模型 / 流程 / 排障 |
  78. | 原则提取器 | `extractors/principle-extractor.md` | **全量扫描**(含公式、表格、脚注) | 原则 / 清单 / 规则 / 计算口径 |
  79. | 案例提取器 | `extractors/case-extractor.md` | 检索式取块(局部命中型) | 书中实例、转述案例、例题,标明类型 |
  80. | 反例提取器 | `extractors/counter-example-extractor.md` | 检索式取块(局部命中型) | 书中警告的失败模式 |
  81. | 术语提取器 | `extractors/glossary-extractor.md` | 检索式取块 + 脚本预筛 | 关键概念词典 |
  82. 每个 sub-agent 独立判断、独立输出到 `books/<slug>/candidates/<type>.md`。
  83. - **长文本**: 超出单个 sub-agent 上下文的内容,按 `methodology/02-stage1-parallel-extract.md` 的分块策略处理;已建立内容索引(`.cangjie/index/`)时,检索式 extractor 按索引取相关块。
  84. - **覆盖率硬门**: 同时对照全量扫描候选和阶段 0 的原书关键任务清单,不只检查旧版通过项。重要任务须有候选及明确去向;有未解释遗漏时补读/退回全量扫描,不虚报零遗漏。
  85. - **降级方案**: 当前环境不支持并行 sub-agent 时,用同样 5 个 extractor prompt **串行**执行,产出格式不变。
  86. ### 阶段 1.5 — 三重验证筛选(知识验证)
  87. 读取 `methodology/03-stage1.5-triple-verify.md`,对每个候选单元执行:
  88. - **V1 来源充分性**: 可定位原文足以支持限定范围内的做法吗?完整机制在一处讲透也可通过。
  89. - **V2 可执行性**: 用合法的新输入能完成任务并核验结果吗?不强求跨领域外推。
  90. - **V3 任务增益**: 能减少遗漏、统一口径或稳定交付吗?不要求作者独创。
  91. 按 `verified / reference / needs_review / rejected` 分流,记录依据及缺口;参考内容须映射到实际交付的 overview/glossary,不是所有未通过项都淘汰。更新 `coverage-audit.md`。
  92. **用户轻确认** ★: 展示可执行候选、参考、待核查、淘汰四类及重要缺口,再进入阶段 1.6;确认不能替代来源验证。
  93. ### 阶段 1.6 — 独立 Skill 晋级门(产品化验证)
  94. 读取 `methodology/03b-stage1.6-promotion-gate.md`。知识验证通过 ≠ 值得成为独立 Skill。对每个单元评审五条判据(独立意图/独立契约/独立运行/独立复用/独立评测,前 3 条必须通过,后 2 条至少 1 条),把去向写入 Bundle 的 `promotion.destination`(promoted / router)。未晋级单元**不淘汰**,保留为来源路由入口内的能力卡。可发现入口软预算默认 8(含 1 个来源路由入口)。
  95. ### 阶段 2 — RIA++ 构造能力卡
  96. 对每个通过的单元,按 `methodology/04-stage2-ria-plus.md` 构造 R / I / A1 / A2 / E / B 六段能力卡:
  97. - 卡片正文写入 `books/<slug>/.cangjie/capabilities/cards/<slug>.md`(**不带 frontmatter**,frontmatter 数据记入 Bundle);
  98. - 同时在 `verified.yaml` 登记该能力: 稳定 `capability_id`、intents、keywords、one_liner、importance(附依据)、`frontmatter.description`(A2 浓缩版)。
  99. - A1 区分原书案例与明确标记的合成演练;E 明确输入、输出、分支和完成标准。必要时通过 `resources` 显式声明配套文本资源,见阶段 2 说明。
  100. ### 阶段 3 — Zettelkasten 链接
  101. 按 `methodology/05-stage3-zettelkasten.md`:
  102. 1. 找出能力之间的引用关系,写入 Bundle 中各能力的 `also_read`,并回填 A2 的"与相邻能力的区分"
  103. 2. 把 `candidates/glossary.md` 整理成 `books/<slug>/GLOSSARY.md`,并复制到 `.cangjie/capabilities/book/glossary.md`
  104. 3. INDEX/路由表由编译器从 Bundle 生成,不再手写
  105. 4. 把 `references.md` 中需保留的内容并入 Bundle 的 `book/overview.md` 或 `book/glossary.md`,在覆盖审计登记实际路径;不把待核查内容伪装为可执行方法
  106. ### 阶段 4 — 压力测试 (darwin 兼容)
  107. 按 `methodology/06-stage4-pressure-test.md`,晋级能力测触发,router 能力测可达;**两类都要实际完成代表任务并核对输出**,不能用“会调用”替代“做得对”。缺失输出计入评测分母,未完成不得声明通过。未过的回炉重做阶段 2。
  108. ### 阶段 5 — 编译与交付
  109. 按 `methodology/07-stage5-deliver.md`:
  110. 1. 生成 `books/<slug>/DIGEST.md` — 面向读者的精华长文
  111. 2. 运行 `python3 scripts/cangjie.py compile --bundle books/<slug>/.cangjie/capabilities --out <目标目录> --output auto`,把决策报告展示给用户轻确认(按推荐 / 改 single / 改 pack)
  112. 3. 询问用户安装位置,把编译产物复制或 symlink 过去
  113. 4. 告知用户: "已完成,可一键喂给 darwin-skill 自动进化"
  114. ## 质量红线 (违反则阻止输出)
  115. 1. 每个能力必须通过**全部**三重验证
  116. 2. 每张能力卡必须有完整的 R / I / A1 / A2 / E / B 六段;A1 无原书案例时如实标注,合成演练不冒充书中事实
  117. 3. 原文引用 ≤150 字/段 (英文 ≤100 词/段)
  118. 4. 每个 active 能力在 `destinations.json` 中恰好一个去向(promoted_to 或 served_by);未晋级能力必须可经来源路由入口到达
  119. 5. 晋级 Skill 的 `description` 必须明确 trigger 条件,并与来源路由入口互有近邻负例
  120. 6. 编译产物必须通过 `scripts/validate_skill_pack.py`(格式/相对引用/frontmatter 100%)
  121. 7. 生成目录只读:检测到本地手改时不得静默覆盖(编译器强制三选一)
  122. 8. 关键任务覆盖有可追溯去向;实际输出评测与来源核查分别记录。存在重要缺口只能交付明确标注范围的草稿,不能宣称完整通过
  123. ## 与 nuwa-skill / darwin-skill 的生态定位
  124. - **nuwa-skill**: 蒸馏人 (思维方式 / 表达 DNA)
  125. - **cangjie-skill** (本 skill): 蒸馏书 (方法论 / 框架 / 原则)
  126. - **darwin-skill**: 进化任意 skill
  127. 三者咬合: 本 skill 输出的评测用例遵循 darwin-skill 格式,以便产出的 skill 可直接接入 darwin 做自动进化。
  128. ## 调用惯例
  129. - **永远先试点 1 本** — 除非用户明确说"批量"
  130. - **阶段之间主动汇报进度** — 不要静默跑完再 dump 结果
  131. - **不凭记忆拆书** — 没文本就停下来问
  132. - **保留审计轨迹** — candidates/ 和 rejected/ 都要留
  133. - **随时可续跑** — 每完成一个阶段就更新 PIPELINE_STATE.md,中断后从状态文件恢复
  134. - **输出策略持久化** — update/repair 默认沿用原输出模式,不因新增材料静默改变产物形态
README.md · README · 283 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. <p align="center">
  2. <a href="./README.zh-CN.md">简体中文</a> ·
  3. <a href="./README.md">English</a> ·
  4. <a href="./README.ja.md">日本語</a>
  5. </p>
  6. <div align="center">
  7. # Cangjie Skill
  8. ### Distill methodologies from books, long-form videos, and podcasts into callable AI Skills
  9. [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](./LICENSE)
  10. [![Version: 2.5.0](https://img.shields.io/badge/Version-2.5.0-8b5cf6.svg)](./CHANGELOG.md)
  11. [![Method: RIA--TV++](https://img.shields.io/badge/Method-RIA--TV++-2ea44f.svg)](./SKILL.md)
  12. [![Platform: OpenClaw](https://img.shields.io/badge/Platform-OpenClaw-1677ff.svg)](https://github.com/openclaw/openclaw)
  13. [![Platform: Claude Code](https://img.shields.io/badge/Platform-Claude%20Code-f97316.svg)](https://code.claude.com/)
  14. [![Platform: DeepSeek Harness](https://img.shields.io/badge/Platform-DeepSeek%20Harness-4f46e5.svg)](#deepseek-harness-plugin)
  15. **Finish reading, watching, or listening—and leave with a methodology you can invoke.**
  16. </div>
  17. ## Cangjie Skill in 66 seconds
  18. https://github.com/user-attachments/assets/87a2cf7b-2114-48aa-aecd-2640a667db40
  19. ## Official Website
  20. 🌐 [Visit the Cangjie Skill official website](https://cangjie-skill.com/)
  21. 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.
  22. ## What's New in v2.5.0
  23. - **Capability Bundle as the single source of truth**: extraction produces stable capability cards and metadata before any installable output is compiled.
  24. - **Two deterministic delivery modes**: compile one router-style Skill (`single`) or a compact pack with a router plus promoted standalone Skills (`pack`).
  25. - **A unified local toolchain**: `scripts/cangjie.py` now covers diagnostics, compilation, output replanning, incremental updates, repair, rollback, evaluation, and benchmarking.
  26. - **Safer evolution**: content-addressed preprocessing, source diffs, impact analysis, transactional patches, edit detection, snapshots, and rollback are included.
  27. - **Registry v2 and website support**: output mode and capability counts are visible without breaking existing Registry v1 entries.
  28. See the [v2.5.0 release notes](./docs/releases/v2.5.0.md) and [changelog](./CHANGELOG.md) for the complete scope and migration notes.
  29. **2026-09-13 refresh (still v2.5.0):** task-first validation now retains complete procedures and formulas explained in a single source location. Output scoring counts missing runs and checks numeric values/units; compiled Skills can carry declared scripts and text templates. [Download the refreshed generic Skill ZIP](https://github.com/kangarooking/cangjie-skill/releases/download/v2.5.0/cangjie-skill-2.5.0.zip) · [SHA256](https://github.com/kangarooking/cangjie-skill/releases/download/v2.5.0/cangjie-skill-2.5.0.zip.sha256). Extract it and install the complete `cangjie-skill/` directory. Existing users must download the refreshed package; check `BUILD_INFO.json` for the source commit and refresh date. The original tag is unchanged, so GitHub's automatic source archives do not contain this refresh.
  30. ## DeepSeek Harness Plugin
  31. cangjie-skill also provides a standalone installation package for DeepSeek Harness. The adapter layer is bundled in the Release package, so no platform-specific wrapper files are added to this repository.
  32. After installing DeepSeek Harness, download the v2.5.0 package and checksum, verify it, then install from the local tarball:
  33. ```bash
  34. mkdir -p ~/.dsh/packages
  35. curl -fL "https://github.com/kangarooking/cangjie-skill/releases/download/v2.5.0/dsh-cangjie-skill-2.5.0.tgz" \
  36. -o ~/.dsh/packages/dsh-cangjie-skill-2.5.0.tgz
  37. curl -fL "https://github.com/kangarooking/cangjie-skill/releases/download/v2.5.0/dsh-cangjie-skill-2.5.0.tgz.sha256" \
  38. -o ~/.dsh/packages/dsh-cangjie-skill-2.5.0.tgz.sha256
  39. (cd ~/.dsh/packages && shasum -a 256 -c dsh-cangjie-skill-2.5.0.tgz.sha256)
  40. dsh plugin --profile web add ~/.dsh/packages/dsh-cangjie-skill-2.5.0.tgz
  41. dsh web
  42. ```
  43. [Download the DeepSeek Harness plugin (for Cangjie Skill v2.5.0)](https://github.com/kangarooking/cangjie-skill/releases/download/v2.5.0/dsh-cangjie-skill-2.5.0.tgz) · [SHA256 checksum](https://github.com/kangarooking/cangjie-skill/releases/download/v2.5.0/dsh-cangjie-skill-2.5.0.tgz.sha256)
  44. After starting a new task, you can say:
  45. ```text
  46. Use cangjie-skill to distill this book into a set of executable Agent Skills: <file path>
  47. ```
  48. ## Why This Exists
  49. There's a recent viral idea: distilling colleagues into AI skills. Even after someone leaves, their experience, tone, and work style can be partially replicated by AI. [nuwa-skill](https://github.com/alchaincyf/nuwa-skill) does exactly this — creating "human skills" like an Elon Musk skill or a Warren Buffett skill. The companion [darwin-skill](https://github.com/alchaincyf/darwin-skill) handles automatic skill evolution.
  50. Distilling people is valuable — nuwa-skill has already proven this. Distilling the content people have **expressed systematically** is a complementary dimension: a book, a long-form interview, a podcast episode, or a long Bilibili or YouTube video can contain methodologies that took the creator years to refine. Rather than imitating someone's expression style, extracting those methodologies and turning them into tools that solve real problems is equally valuable.
  51. There's also a real pain point: you may read many books, save many videos, and listen to many podcasts, yet still struggle to apply what you learned. Content-rich long videos are published every day, are often time-sensitive, and can be difficult to absorb in one viewing; they may not be represented in an AI model's training data at all. Once this content is distilled into skills, an AI agent can invoke the knowledge in real scenarios instead of letting it gather dust in notes, bookmarks, or watch-later lists.
  52. So cangjie-skill has one clear goal: **distill every piece of high-value content worth distilling**. It works not only with books, but also with videos that have subtitles or transcripts, podcasts, interviews, talks, courses, long-form articles, and document collections. Whenever content contains extractable, verifiable, and transferable methodologies, cangjie-skill can turn them into independently callable, composable, and pressure-testable AI skill packs.
  53. For video content, we recommend using the [video-downloader](https://github.com/kangarooking/kangarooking-skills/tree/main/video-downloader) skill alongside cangjie-skill. Use it first to download the video, extract subtitles or audio transcripts, and collect key materials; then pass the resulting text to cangjie-skill for methodology extraction, skill construction, and pressure testing.
  54. ## What Problems It Solves
  55. - Reading many books, watching many videos, or listening to many podcasts without applying them — knowledge stays at the "read/watched/listened/saved" level and cannot be invoked in real decisions
  56. - Summaries, notes, and organized transcripts are compression, not structured reuse — after reading or watching, you still do not know "what to use when"
  57. - Only a small fraction of high-value content deserves to become a tool — strict filtering is needed, not wholesale inclusion
  58. - Existing methods for reading, watching, and learning are designed for people, not agents — distillation must be execution-oriented rather than consumption-oriented
  59. ## How It Works
  60. cangjie-skill uses the **RIA-TV++** pipeline to transform source texts—including books, video transcripts, podcast transcripts, and interview notes—into a reusable Capability Bundle, then compiles that source into installable skills. The process has seven stages:
  61. > **Script dependencies**: the deterministic scripts under `scripts/` require **Python 3.10+** and **PyYAML** (`python3 -m pip install pyyaml`). `python3 scripts/cangjie.py doctor` runs a self-check and works even when PyYAML is missing. Optional: `tiktoken`, `jsonschema`.
  62. 1. **Whole-Content Comprehension (Adler Analysis)** — Structural, interpretive, critical, and applicability analysis using Mortimer Adler's method, producing `BOOK_OVERVIEW.md`
  63. 2. **Parallel Extraction** — Five specialized extractors (frameworks, principles, cases, counter-examples, glossary) run simultaneously to pull candidate units from the source text
  64. 3. **Triple Verification + Promotion Gate** — Check source sufficiency, executability, and task utility by candidate type. A complete procedure or formula explained once can qualify; repetition or author originality is not mandatory. References and unresolved candidates remain auditable, and standalone entrypoints are decided separately
  65. 4. **RIA++ Capability Construction** — Verified content is structured into R / I / A1 / A2 / E / B capability cards inside `.cangjie/capabilities/`
  66. 5. **Zettelkasten Linking** — Dependencies, contrasts, and compositions are encoded in the Bundle's capability graph and shared glossary
  67. 6. **Pressure Testing** — Test prompts including bait questions (and cross-skill confusion tests) are designed for each skill; failures go back for full reconstruction
  68. 7. **Deterministic Compilation and Delivery** — The same Bundle compiles to `single` or compact `pack`, alongside a reader-facing `DIGEST.md`, validation results, and installable artifacts
  69. The name RIA-TV++ breaks down as:
  70. - **RIA**: From Zhao Zhou's bookmark method (Reading / Interpretation / Appropriation)
  71. - **TV**: Triple Verification
  72. - **++**: Agent-oriented extensions — E (Execution) + B (Boundary)
  73. ## Effect Examples
  74. ### Example 1: From a Book or Long-Form Video to a Skill Toolkit
  75. **User Need**
  76. "I want to turn the core methodologies from a book or a long Bilibili/YouTube video into reusable AI skills, not just a summary."
  77. **How cangjie-skill reasons**
  78. - Check whether the source material has reusable methodological units
  79. - Distinguish what deserves to be a standalone skill vs. background material
  80. - Output a structured skill repository, not a single summary document
  81. **Example Output**
  82. > The result will not be one summary document. It will be a multi-skill repository with `BOOK_OVERVIEW.md`, `INDEX.md`, a reader-facing `DIGEST.md`, a `GLOSSARY.md`, multiple `*/SKILL.md` files, and `test-prompts.json` for trigger testing.
  83. ### Example 2: Structured Reuse, Not Compression
  84. **User Need**
  85. "I don't want a long explanatory article. I want a skill pack my agent can reuse."
  86. **How cangjie-skill reasons**
  87. - Target is structured reuse, not narrative compression
  88. - Prioritize triggerable, composable, testable skill units
  89. - Reject material that doesn't deserve standalone skill status
  90. **Example Output**
  91. > The system produces multiple skill modules with trigger conditions, boundaries, execution patterns, and related-skill links — rather than flattening the source into one generalized note.
  92. ## Generated Skill Packs
  93. | Repository | Source | Skills |
  94. |------------|--------|--------|
  95. | [buffett-letters-skill](https://github.com/kangarooking/buffett-letters-skill) | Buffett's shareholder letters (1957-2023) | 20 |
  96. | [cognitive-dividend-skill](https://github.com/kangarooking/cognitive-dividend-skill) | Cognitive Dividend | 15 |
  97. | [duan-yongping-skill](https://github.com/kangarooking/duan-yongping-skill) | Duan Yongping's Q&A (business + investment logic) | 15 |
  98. | [viral-copywriting-skill](https://github.com/kangarooking/viral-copywriting-skill) | Bao Kuan Wen An | 14 |
  99. | [copywriters-handbook-skill](https://github.com/kangarooking/copywriters-handbook-skill) | The Copywriter's Handbook | 12 |
  100. | [contagious-skill](https://github.com/kangarooking/contagious-skill) | Contagious | 15 |
  101. | [influence-skill](https://github.com/kangarooking/influence-skill) | Influence | 12 |
  102. | [1000-true-fans-skill](https://github.com/kangarooking/1000-true-fans-skill) | 1000 True Fans | 13 |
  103. | [system-prompt-skills](https://github.com/kangarooking/system-prompt-skills) | 165 AI product system prompts | 15 |
  104. | [X-growth-skills](https://github.com/kangarooking/X-growth-skills) | Practical X (Twitter) account launch, content growth, algorithm, engagement, and monetization resources | 15 |
  105. | [sunyuchen-skill](https://github.com/kangarooking/sunyuchen-skill) | A single narrative writing sample labeled “sunyuchen” | 1 (7 capabilities) |
  106. | [poor-charlies-almanack-skill](https://github.com/kangarooking/poor-charlies-almanack-skill) | Poor Charlie's Almanack | 12 |
  107. | [no-rules-rules-skill](https://github.com/kangarooking/no-rules-rules-skill) | No Rules Rules | 10 |
  108. | [huangdi-neijing-skill](https://github.com/kangarooking/huangdi-neijing-skill) | *Huangdi Neijing* (*Suwen* + *Lingshu*) | 22 |
  109. | [first-principles-skill](https://github.com/kangarooking/first-principles-skill) | First Principles | 10 |
  110. | [mao-selected-works-skill](https://github.com/kangarooking/mao-selected-works-skill) | Selected Works of Mao Zedong, Vol. 1-5 | 25 |
  111. | [qbdx-hub/buffett-letters-skill](https://github.com/qbdx-hub/buffett-letters-skill) | Buffett Shareholder Letters (1957-2023) | 20 |
  112. | [qbdx-hub/wo-yu-di-tan-skill](https://github.com/qbdx-hub/wo-yu-di-tan-skill) | Wo Yu Di Tan | 6 |
  113. | [qbdx-hub/mingchao-those-things-skill](https://github.com/qbdx-hub/mingchao-those-things-skill) | Mingchao Those Things | 7 |
  114. | [qbdx-hub/sunzi-bingfa-skill](https://github.com/qbdx-hub/sunzi-bingfa-skill) | Sunzi Bingfa | 8 |
  115. | [qbdx-hub/zhouyi-skill](https://github.com/qbdx-hub/zhouyi-skill) | Zhouyi | 8 |
  116. | [qbdx-hub/high-math-vol1-ch1-skill](https://github.com/qbdx-hub/high-math-vol1-ch1-skill) | High Math Vol. 1 Chapter 1 | 8 |
  117. ## Video Distillation
  118. These repositories are built from subtitles or transcripts of long-form videos, courses, or video collections. They demonstrate cangjie-skill's ability to distill methodologies from non-book content.
  119. | Repository | Source | Skills |
  120. |------------|--------|--------|
  121. | [ai-for-everyone-skill](https://github.com/kangarooking/ai-for-everyone-skill) | Andrew Ng's *AI for Everyone* video course | 25 |
  122. | [loop-engineering-skill](https://github.com/kangarooking/loop-engineering-skill) | Loop Engineering long-form video collection | 8 |
  123. More high-value books are planned for distillation. Future candidates include, but are not limited to, *The Prince*.
  124. Additional external source (included with the author's permission):
  125. - Source repository: [ace3000chao/book2startup](https://github.com/ace3000chao/book2startup)
  126. - Included books: *The Lean Startup*, *The Art of War*, *Zhuangzi*, and *I Ching*
  127. - Source repository: [shenqistart/book2skill](https://github.com/shenqistart/book2skill)
  128. - Included books: *Chanlun* and *The Classic of Tea*
  129. ## Repository Structure
  130. ```text
  131. cangjie-skill/
  132. ├── README.md ← You are here (default)
  133. ├── README.zh-CN.md ← Simplified Chinese version
  134. ├── README.ja.md ← Japanese version
  135. ├── LICENSE ← MIT License
  136. ├── SKILL.md ← Meta-skill definition (full execution spec for cangjie-skill)
  137. ├── methodology/ ← RIA-TV++ stage-by-stage methodology docs
  138. ├── extractors/ ← Prompt definitions for the 5 parallel extractors
  139. └── templates/ ← SKILL.md / INDEX.md / BOOK_OVERVIEW.md templates
  140. ```
  141. ## Ecosystem
  142. cangjie-skill is part of a larger skill ecosystem:
  143. - [nuwa-skill](https://github.com/alchaincyf/nuwa-skill) — Distills people (thinking styles, expression DNA)
  144. - **cangjie-skill** (this repo) — Distills books (methodologies, frameworks, principles)
  145. - [darwin-skill](https://github.com/alchaincyf/darwin-skill) — Evolves any skill
  146. They interlock: nuwa distills people, cangjie distills books, darwin keeps them evolving.
  147. ## More Skills
  148. - [Buffett Letters Skill](https://github.com/kangarooking/buffett-letters-skill) — 20 investment reasoning skills from Buffett's 60+ years of shareholder letters
  149. - [Poor Charlie's Almanack Skill](https://github.com/kangarooking/poor-charlies-almanack-skill) — 12 decision-making and judgment skills from Charlie Munger's core thinking methods
  150. - [No Rules Rules Skill](https://github.com/kangarooking/no-rules-rules-skill) — 10 organizational design skills from Netflix's culture of freedom and responsibility
  151. - [Cognitive Dividend Skill](https://github.com/kangarooking/cognitive-dividend-skill) — 15 cognitive tool skills for thinking upgrades from Cognitive Dividend
  152. - [Duan Yongping Skill](https://github.com/kangarooking/duan-yongping-skill) — 15 business and investment skills from Duan Yongping's Q&A collection
  153. - [Viral Copywriting Skill](https://github.com/kangarooking/viral-copywriting-skill) — 14 sales copywriting and diagnosis skills from *Bao Kuan Wen An*
  154. - [Copywriters Handbook Skill](https://github.com/kangarooking/copywriters-handbook-skill) — 12 sales copywriting, headline, and benefit translation skills from *The Copywriter's Handbook*
  155. - [Contagious Skill](https://github.com/kangarooking/contagious-skill) — 15 STEPPS propagation strategy and word-of-mouth diagnosis skills from *Contagious*
  156. - [Influence Skill](https://github.com/kangarooking/influence-skill) — 12 persuasion psychology, compliance mechanism, and defensive judgment skills from *Influence*
  157. - [1000 True Fans Skill](https://github.com/kangarooking/1000-true-fans-skill) — 13 personal branding, true fan development, and trust-based monetization skills from *1000 True Fans*
  158. - [System Prompt Skills](https://github.com/kangarooking/system-prompt-skills) — 15 system prompt design skills distilled from 165 AI product system prompts
  159. - [X Growth Skills](https://github.com/kangarooking/X-growth-skills) — 15 skills for X account launch, content, algorithms, engagement, review, and monetization
  160. - [sunyuchen-skill](https://github.com/kangarooking/sunyuchen-skill) — A restrained narrative writing skill covering cold opens, operational detail, short dialogue, emotional restraint, and object callbacks
  161. - [Huangdi Neijing Skill](https://github.com/kangarooking/huangdi-neijing-skill) — 22 methodology skills from *Huangdi Neijing*, including 12 from *Suwen* and 10 from *Lingshu*
  162. - [First Principles Skill](https://github.com/kangarooking/first-principles-skill) — 10 skills on axiomatic reasoning, boundary-breaking innovation, and organizational refresh from *First Principles*
  163. - [Mao Selected Works Skill](https://github.com/kangarooking/mao-selected-works-skill) — 25 cognition, strategy, organization, and execution skills from *Selected Works of Mao Zedong*
  164. - [qbdx-hub Buffett Letters Skill](https://github.com/qbdx-hub/buffett-letters-skill) — 20 investment and capital allocation skills from Buffett shareholder letters
  165. - [qbdx-hub Wo Yu Di Tan Skill](https://github.com/qbdx-hub/wo-yu-di-tan-skill) — 6 skills on limits, suffering, writing, and self-anchoring from *Wo Yu Di Tan*
  166. - [qbdx-hub Mingchao Those Things Skill](https://github.com/qbdx-hub/mingchao-those-things-skill) — 7 skills on power structure, institutional failure, and historical explanation from *Mingchao Those Things*
  167. - [qbdx-hub Sunzi Bingfa Skill](https://github.com/qbdx-hub/sunzi-bingfa-skill) — 8 skills on strategic judgment, resource control, and action selection from *Sunzi Bingfa*
  168. - [qbdx-hub Zhouyi Skill](https://github.com/qbdx-hub/zhouyi-skill) — 8 skills on situational diagnosis, timing, and advance-retreat boundaries from *Zhouyi*
  169. - [qbdx-hub High Math Vol. 1 Chapter 1 Skill](https://github.com/qbdx-hub/high-math-vol1-ch1-skill) — 8 learning skills on limits, infinitesimals, and continuity from High Math Vol. 1 Chapter 1
  170. External Source (included with the author's permission):
  171. - [book2startup](https://github.com/ace3000chao/book2startup) — includes skills distilled from *The Lean Startup*, *The Art of War*, *Zhuangzi*, and *I Ching*
  172. - [book2skill](https://github.com/shenqistart/book2skill) — includes AI-Agent skills distilled from *Chanlun* and *The Classic of Tea*
  173. ## Contributors
  174. Thank you to the following contributors for expanding the cangjie-skill ecosystem:
  175. - [shenqistart](https://github.com/shenqistart) — contributed the external [book2skill](https://github.com/shenqistart/book2skill) reference and additions across the Chinese, English, and Japanese READMEs
  176. - [qbdx-hub](https://github.com/qbdx-hub) — contributed 6 Cangjie whole-book/chapter distillation example repositories and additions across the Chinese, English, and Japanese READMEs
  177. ## About the Author
  178. **袋鼠帝 kangarooking** — AI blogger and indie developer. Creator of the AI Top WeChat Official Account “袋鼠帝 AI 客栈”
  179. <img src="https://raw.githubusercontent.com/kangarooking/cangjie-skill/main/assets/wechat-personal-qr.jpg" width="220" alt="Kangarooking personal WeChat QR code">
  180. Volcengine Navigation KOL, Baidu Qianfan Developer Ambassador, GLM Evangelist, Trae Kunming's First Fellow
  181. | Platform | Link |
  182. |----------|------|
  183. | 𝕏 Twitter | https://x.com/aikangarooking |
  184. | Xiaohongshu | https://xhslink.com/m/5YejKvIDBbL |
  185. | Douyin | https://v.douyin.com/hYpsjphuuKc |
  186. | WeChat Official Account | 袋鼠帝 AI 客栈 |
  187. | WeChat Video Channel | AI 袋鼠帝 |
  188. WeChat Official Account「袋鼠帝 AI 客栈」QR code:
  189. ![](https://raw.githubusercontent.com/kangarooking/cangjie-skill/main/assets/kangarooking-gzh.png)
  190. If you also want to distill methodologies from books, long-form videos, podcasts, and courses into callable Agent Skills, join the cangjie-skill WeCom community group:
  191. <img src="https://raw.githubusercontent.com/kangarooking/cangjie-skill/main/assets/wecom-cangjie-group-qr.png" width="220" alt="cangjie-skill WeCom community group QR code">
  192. ## ⭐ Star History
  193. If this project has helped you, please star it.
  194. <a href="https://www.star-history.com/?repos=kangarooking%2Fcangjie-skill&type=date&legend=top-left">
  195. <img alt="Star History Chart" src="./assets/star-history.svg" />
  196. </a>
  197. ## License
  198. MIT License. See [LICENSE](./LICENSE).

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.