book-reader
由 josharsh 提供
PDF & Documents简介
Read books (epub, pdf, txt) from various sources with progress tracking.
描述
该技能的功能
Lets AI agents read full-length books in EPUB, PDF, and TXT formats from local files or online sources. Tracks reading progress across sessions so agents can resume where they left off. Supports chunked reading and key insight extraction.
Persistent progress tracking and multi-format support let agents work through book-length content without losing context between sessions, something most document tools don't handle.
何时使用
- Summarizing a non-fiction book chapter by chapter
- Extracting key arguments from a downloaded PDF paper
- Resuming a long book read across multiple agent sessions
- Searching Project Gutenberg for public domain texts to analyze
- Building a reading list and tracking progress across titles
示例工作流
User asks: Read the first 50 pages of Thinking, Fast and Slow and summarize the main ideas
- 1
Search for the book using book-reader.sh search
- 2
Download the EPUB from the matched source
- 3
Read the first 50 pages with --pages 50 flag
- 4
Extract key insights during chunked reading
- 5
Save progress to reading-state.json for later sessions
A summary of the opening chapters covering System 1 vs System 2 thinking, with bookmarks and notes saved for the next session
前置要求
- poppler-utils installed (pdftotext for PDF support)
- pandoc installed (for EPUB conversion, optional)
- Python 3 with ebooklib, beautifulsoup4, and lxml packages
安装
clawhub install josharsh/book-reader运行 CLI 命令,或把这段提示词粘贴到 Claude Code、Cursor、Codex 或你在用的任意编程智能体里。