File structure
CODEBLOCK0
What it tracks
In progress: What you're currently reading, where you are, any notes mid-read.
Finished: Rating, key ideas, best quotes, what changed your thinking.
Abandoned: Books you gave up on and why — this data is useful too.
Want to read: Recommendations with context (who recommended it, why it matters to you).
Adding books
Start reading: /book start [title] [author]
Finish reading: /book done [title] — triggers a brief reflection prompt
Add to list: /book want [title] — adds with optional reason
Log a note mid-read: /book note [title] [note]
Abandoned: /book drop [title] [why]
Post-reading reflection
When you mark a book as done, the agent asks 4 questions:
- 1. What's the one idea from this book you'll actually use or remember?
- What surprised you?
- Who would you recommend this to and why?
- Rating: 1-5
Short answers are fine. One sentence each is enough.
These become the book's entry in library.md.
library.md structure
CODEBLOCK1
Reading list intelligence
INLINECODE5 — suggests what to read next based on:
- - What you've enjoyed most (by rating and genre)
- Themes you've been exploring lately
- Books you've had on the list longest
- Any books that connect to what you're working on or thinking about
Not just a ranked list. A recommendation with a reason:
CODEBLOCK2
Theme tracking
After 10+ books, patterns become visible.
INLINECODE6 — what have you been reading about?
CODEBLOCK3
Book and knowledge-capture integration
When you log a quote or key idea from a book:
INLINECODE7
The skill offers to add it to knowledge-capture:
"Add this to your knowledge base? It'll be searchable there alongside your other notes."
The two skills build on each other. Books feed the knowledge base.
The knowledge base surfaces relevant reading notes when you're thinking about a topic.
Reading goal
Optional. Set a books-per-month target if you want one.
INLINECODE8
At the end of each month, a brief note:
"Read 2 books in March, goal was 3. On track for the year (8 of 10)."
Not a guilt trip. Just a number.
Privacy rules
Reading history is personal and low-stakes, but still private.
Context boundary: Only surface reading lists and notes in the owner's private channel.
Never share what someone is reading or their book notes in a group context.
Approval gate: No book is added to or removed from the library without
the owner's instruction — either explicit command or confirmation of a suggestion.
Management commands
- -
/book start [title] [author] — start reading - INLINECODE10 — mark finished, trigger reflection
- INLINECODE11 — mark abandoned
- INLINECODE12 — add to reading list
- INLINECODE13 — log a note
- INLINECODE14 — get a recommendation
- INLINECODE15 — show reading list
- INLINECODE16 — show all books
- INLINECODE17 — reading pattern analysis
- INLINECODE18 — search your library
- INLINECODE19 — show entry for one book
文件结构
book-tracker/
SKILL.md
library.md ← 每本书的状态、笔记、评分
reading-list.md ← 想读的书及理由
config.md ← 类型、偏好、月度目标
追踪内容
阅读中: 你当前在读的书、读到哪了、阅读过程中的笔记。
已读完: 评分、核心观点、最佳引文、改变了你哪些想法。
弃读: 你放弃的书及原因——这些数据同样有用。
想读: 带有背景信息的推荐(谁推荐的、为什么对你有意义)。
添加书籍
开始阅读: /book start [书名] [作者]
读完: /book done [书名] — 触发简短的反思提示
加入列表: /book want [书名] — 可附带理由
阅读中记录笔记: /book note [书名] [笔记]
弃读: /book drop [书名] [原因]
读后反思
当你标记一本书为已读完时,助手会问 4 个问题:
- 1. 这本书中你实际会使用或记住的一个观点是什么?
- 什么让你感到惊讶?
- 你会向谁推荐这本书,为什么?
- 评分:1-5 分
简短回答即可。每问一句话就够。
这些内容会成为该书在 library.md 中的条目。
library.md 结构
md
[书名] — [作者]
状态:已读 / 阅读中 / 弃读 / 想读
读完日期:[日期]
评分:[1-5]
类型:[小说 / 非虚构 / 传记 / 等]
核心观点:[用户反思中的回答]
感到惊讶:[用户的回答]
推荐给:[用户的回答]
笔记:
[任何阅读中或读后的笔记]
引文:
[任何值得保留的段落]
关联书籍:[图书馆中相关的其他书籍]
阅读列表智能推荐
/book next — 根据以下因素推荐下一本读物:
- - 你最喜欢的书籍(按评分和类型)
- 你最近在探索的主题
- 在你的列表中最久的书籍
- 与你当前工作或思考相关的书籍
不只是排名列表,而是带有理由的推荐:
📖 接下来读什么
- 1. 《妈妈测试》 — 罗布·菲茨帕特里克
为什么现在读:你一直在阅读关于产品和客户研究的内容。
这是列表中对你当前工作最实用的一本书。
在你的列表上已有:4个月([姓名]推荐时添加)
- 2. 《皮拉内西》 — 苏珊娜·克拉克
为什么现在读:你已经连续3个月阅读大量非虚构作品。
你的评分显示你喜欢用文学小说来调剂。
篇幅短小。风格独特。不会花太长时间。
主题追踪
读完 10 本书后,模式会变得明显。
/book themes — 你一直在读什么?
📖 阅读主题 — 过去 12 个月
最频繁的主题:
- 1. 产品与战略(8本书)
- 管理与领导力(5本书)
- 文学小说(4本书)
- 历史(3本书)
你可能注意到的空白:
• 你已经 18 个月没怎么读科学或技术类书籍了
• 你的小说与非虚构比例是 1:4(大量非虚构阅读)
你评分最高的书籍往往是:[类型/主题模式]
书籍与知识捕捉整合
当你记录书中的引文或核心观点时:
/book note [书名] 这句话值得保留
该技能会提供选项将其加入知识库:
要将其添加到你的知识库吗?它将在那里与你的其他笔记一起可搜索。
两项技能相辅相成。书籍为知识库提供素材。
当你在思考某个主题时,知识库会呈现相关的阅读笔记。
阅读目标
可选。如果你愿意,可以设定每月阅读目标。
/book goal 每月 [N] 本
每月月底会有一条简短提示:
3月读了2本书,目标是3本。年度进度正常(已完成8/10)。
不是施加压力,只是一个数字。
隐私规则
阅读历史是个人化的,风险较低,但仍然是私密的。
上下文边界: 仅在拥有者的私人频道中展示阅读列表和笔记。
绝不在群组环境中分享某人正在读什么或他们的读书笔记。
批准机制: 未经拥有者指示——无论是明确指令还是对建议的确认——不得向图书馆添加或移除任何书籍。
管理命令
- - /book start [书名] [作者] — 开始阅读
- /book done [书名] — 标记为已读完,触发反思
- /book drop [书名] — 标记为弃读
- /book want [书名] — 加入阅读列表
- /book note [书名] [笔记] — 记录笔记
- /book next — 获取推荐
- /book list — 显示阅读列表
- /book library — 显示所有书籍
- /book themes — 阅读模式分析
- /book search [查询] — 搜索你的图书馆
- /book [书名] — 显示单本书的条目