By name: book-reading-assistant description: This skill assists with reading technical books through chapter-by-chapter analysis, comprehension testing, and persistent note-taking. Use this skill when the user wants to read and deeply understand a technical book (PDF/EPUB format), needs structured reading assistance across multiple sessions, or wants to track progress and maintain organized reading notes. Triggers include requests to start reading a book, analyze chapters, save reading notes, track terminology, or schedule review sessions.
Book Reading Assistant
Overview
This skill enables systematic reading and comprehension of technical books through chapter-by-chapter analysis, interactive learning assistance, and persistent note-taking. Designed for multi-session learning workflows, it helps readers extract key insights, validate understanding, and maintain organized knowledge bases from technical literature.
The skill operates in three reading stages (pre-reading, during-reading, post-reading) and maintains persistent files for continuous learning across sessions.
Book Initialization Workflow
When the user wants to start reading a new book, follow this initialization sequence:
1. Request Book File and Output Directory
Ask the user to provide:
- Path to the book file (PDF or EPUB)
- Desired output directory for persistent notes
2. Extract Table of Contents
Use the Read tool to load the book file and identify the table of contents:
- For PDFs: Look for TOC in the first 10-20 pages
- For EPUBs: Claude's Read tool will display the structure
- Identify chapter titles, section headings, and page ranges
- Extract book metadata (title, author, publication info if available)
3. Create Directory Structure
Initialize the output directory with this structure:
<output-directory>/
├── book-metadata.json
├── chapters/
├── glossary.md
├── cross-chapter-analysis.md
└── review-schedule.json
4. Initialize book-metadata.json
Create the book metadata file with this structure:
{
"title": "Book Title",
"author": "Author Name",
"file_path": "/path/to/book.pdf",
"total_chapters": 12,
"table_of_contents": [
{
"chapter": 1,
"title": "Chapter Title",
"page_start": 15,
"page_end": 42,
"status": "not_started",
"last_accessed": null,
"comprehension_score": null
}
],
"reading_progress": {
"chapters_completed": 0,
"current_chapter": null,
"last_session": null,
"total_reading_time_minutes": 0
},
"created_at": "2025-01-20T10:30:00Z"
}
5. Initialize Supporting Files
Create empty placeholder files:
glossary.md: "# Technical Glossary\n\nTerms will be added as you read"cross-chapter-analysis.md: "# Cross-Chapter Analysis\n\n## Recurring Themes\n\n## Conceptual Connections\n\n## Evolution of Ideas"review-schedule.json:{"reviews": []}
6. Confirm Initialization
Inform the user that the book has been initialized and show:
- Total chapters detected
- Book title and author
- Output directory location
- Suggested next action: "Ready to begin. Which chapter would you like to start with?"
Chapter Analysis Workflow
When the user indicates they want to analyze a chapter (e.g., "I'm about to read Chapter 3" or "Let's analyze Chapter 1"):
1. Load Chapter Content
- Read the book file using the Read tool, focusing on the specific page range for the chapter
- If the chapter is long (>30 pages), consider processing it in sections
2. Determine Reading Stage
Identify what stage of reading assistance is needed:
- Pre-reading: User says "about to read", "preview", "what should I focus on"
- During-reading: User says "I'm reading", "expand on this", "explain more about"
- Post-reading: User says "finished", "completed", "test my understanding"
3. Apply Stage-Specific Workflow
Execute the appropriate reading stage workflow (detailed in next section)
4. Update Persistence Files
After each interaction, update the relevant files:
- Chapter note file (if analysis was performed)
- book-metadata.json (update status, timestamps)
- glossary.md (add new terms)
- cross-chapter-analysis.md (if patterns emerge)
- review-schedule.json (if chapter was completed)
Reading Stage Workflows
Pre-Reading Stage
Goal: Prepare the reader for effective chapter comprehension by previewing key concepts.
Process:
- Quickly scan the chapter content
- Identify 3-5 core concepts or topics covered
- Highlight technical terms that will appear
- Note any prerequisite knowledge from previous chapters
- Provide a brief roadmap of the chapter structure
Output format:
## Chapter X Pre-Reading Brief
### Focus Areas
1. [Key concept 1] - This section introduces...
2. [Key concept 2] - Pay attention to...
3. [Key concept 3] - Understanding this will help...
### Key Terms to Watch For
- **Term 1**: You'll encounter this in the context of...
- **Term 2**: This builds on the concept from Chapter Y...
### Chapter Roadmap
- Introduction (pages X-Y): Sets up the problem
- Core Content (pages Y-Z): Presents three main solutions
- Examples (pages Z-W): Practical applications
### Prerequisite Connections
This chapter builds on concepts from:
- Chapter X: [concept]
- Chapter Y: [concept]
During-Reading Stage
Goal: Enhance comprehension through content expansion and divergent thinking.
Process:
- When user mentions a specific concept or section, provide deeper context
- Offer alternative perspectives or analogies
- Connect to real-world applications
- Relate to other chapters or external knowledge
- Anticipate potential confusion points and clarify
Interaction style:
- Wait for user prompts like "expand on X" or "tell me more about Y"
- Provide concise, focused explanations
- Use analogies and examples relevant to technical readers
- Ask probing questions to deepen thinking: "How might this apply to...?"
Post-Reading Stage
Goal: Validate comprehension, create comprehensive chapter notes, and schedule reviews.
Process:
1. Create Full Chapter Analysis
Generate a comprehensive chapter note file using the 9-section format (see Chapter Note Format section below). Save this as chapters/chapter-XX-[title-slug].md
2. Interactive Comprehension Testing
Conduct a comprehension check dialogue:
## Comprehension Check - Chapter X
Let's verify your understanding of the key concepts.
### Question 1: [Core concept]
[Thought-provoking question requiring synthesis]
*[After user answers, provide feedback]*
✓ Correct! You've grasped...
or
↻ Not quite. The chapter emphasized that...
### Question 2: [Application]
[Question about practical application]
### Question 3: [Connection]
How does [concept X] relate to [concept Y] from Chapter Z?
Ask 3-5 questions based on the chapter's core teachings. Evaluate user responses and provide constructive feedback.
3. Calculate Comprehension Score
Based on the user's answers, assign a comprehension score (0-100):
- 90-100: Excellent understanding, minimal review needed
- 70-89: Good grasp, schedule one review
- 50-69: Partial understanding, schedule two reviews
- <50: Needs re-reading, schedule intensive review
4. Update Progress Tracking
Update book-metadata.json:
- Set chapter status to "completed"
- Record comprehension_score
- Update chapters_completed count
- Set last_accessed timestamp
5. Schedule Spaced Repetition Review
Add review entries to review-schedule.json based on the spaced repetition algorithm (see Spaced Repetition section)
6. Update Glossary
Extract new technical terms from the chapter and add to glossary.md with definitions
7. Update Cross-Chapter Analysis
If this chapter introduces themes or concepts that connect to previous chapters, update cross-chapter-analysis.md
Chapter Note Format
Each chapter analysis file (chapters/chapter-XX-[title-slug].md) follows this 9-section structure:
# Chapter X: [Chapter Title]
## 1. Chapter Metadata
- **Chapter Number**: X
- **Chapter Title**: [Title]
- **Page Range**: XX-YY
- **Date Completed**: YYYY-MM-DD
- **Comprehension Score**: XX/100
## 2. Key Quotes
> "Quote 1 - most powerful or central idea from the chapter"
> *(Page XX)*
> "Quote 2 - thought-provoking statement"
> *(Page YY)*
[4-8 quotes total - capture verbatim with page numbers]
## 3. Main Stories / Examples
**Example 1: [Title]**
- [Brief summary of story/anecdote]
- **Moral/Meaning**: [Key takeaway]
**Example 2: [Title]**
- [Summary]
- **Moral/Meaning**: [Takeaway]
## 4. Chapter Summary
[A clear, concise paragraph (4-6 sentences) summarizing the entire chapter. Capture the main arc: what problem is introduced, what solutions/concepts are presented, and what conclusions are reached.]
## 5. Core Teachings
The main ideas, arguments, or lessons from this chapter:
1. **[Teaching 1]**: [Explanation]
2. **[Teaching 2]**: [Explanation]
3. **[Teaching 3]**: [Explanation]
[Continue as needed]
## 6. Actionable Lessons
Practical lessons or advice that can be applied:
- **[Lesson 1]**: [How to apply]
- **[Lesson 2]**: [How to apply]
- **[Lesson 3]**: [How to apply]
## 7. Mindset / Philosophical Insights
Deeper reflections, shifts in thinking, or philosophical takeaways:
- [Insight 1]
- [Insight 2]
- [Insight 3]
## 8. Memorable Metaphors & Analogies
**[Metaphor 1]**: [Original metaphor from the book]
- **Meaning**: [What it illustrates]
**[Metaphor 2]**: [Another comparison]
- **Meaning**: [Explanation]
## 9. Questions for Reflection
1. [Thought-provoking question related to core concept]
2. [Question about application or implications]
3. [Question connecting to broader context]
4. [Question encouraging critical thinking]
5. [Question about personal application]
---
## Reading Notes
[Optional section for any additional observations, personal insights, or questions that arose during reading]
Writing Guidelines:
- Capture quotes verbatim - exact wording from the book
- All other content should be in clear, engaging language
- Keep explanations free of unnecessary fluff
- Maintain the original context and tone of the book
- Make notes human-readable and suitable for future review
Cross-Chapter Analysis
As the user progresses through multiple chapters, maintain cross-chapter-analysis.md to track:
Recurring Themes
Identify concepts, principles, or topics that appear across multiple chapters:
## Recurring Themes
### Theme: [Name]
- **Chapters**: 1, 3, 5, 7
- **Evolution**: How this theme develops across chapters
- **Key Insight**: What the recurring emphasis reveals
### Theme: [Name]
- **Chapters**: 2, 4, 6
- **Evolution**: ...
Conceptual Connections
Map relationships between concepts from different chapters:
## Conceptual Connections
### Connection: [Concept A] ↔ [Concept B]
- **Concept A** (Chapter X): [Brief description]
- **Concept B** (Chapter Y): [Brief description]
- **Relationship**: How they relate, dependencies, or contrast
Evolution of Ideas
Track how the author builds arguments progressively:
## Evolution of Ideas
### Idea: [Core Argument]
- **Chapter 1**: Foundation - [What's established]
- **Chapter 3**: Expansion - [How it's built upon]
- **Chapter 5**: Application - [Practical implementation]
- **Chapter 8**: Advanced - [Complex extensions]
Update frequency: Add to cross-chapter analysis after completing each chapter, but only when genuine connections exist. Not every chapter will contribute to every theme.
Terminology Glossary Management
Maintain glossary.md as an alphabetically organized reference of technical terms:
# Technical Glossary
## A
**[Term]**
- **Definition**: [Clear, concise explanation]
- **Chapter Introduced**: X
- **Context**: [How it's used in the book]
- **Related Terms**: [term1], [term2]
## B
**[Term]**
- **Definition**: ...
Process:
- When processing a chapter, identify technical terms that are:
- Newly introduced concepts
- Domain-specific terminology
- Terms with specific meaning in the book's context
- Add 5-10 most important terms per chapter
- Maintain alphabetical order
- Cross-reference related terms
- Include the chapter where each term first appears
Spaced Repetition System
Schedule chapter reviews using a spaced repetition algorithm to combat the forgetting curve.
Review Intervals
Base review schedule on comprehension score:
High comprehension (90-100):
- Review 1: +7 days
- Review 2: +30 days
- Review 3: +90 days
Good comprehension (70-89):
- Review 1: +3 days
- Review 2: +14 days
- Review 3: +45 days
- Review 4: +90 days
Moderate comprehension (50-69):
- Review 1: +1 day
- Review 2: +7 days
- Review 3: +21 days
- Review 4: +60 days
Low comprehension (<50):
- Suggest immediate re-reading
- Review 1: +2 days (after re-reading)
- Review 2: +7 days
- Review 3: +21 days
review-schedule.json Format
{
"reviews": [
{
"chapter": 1,
"chapter_title": "Introduction to Neural Networks",
"completed_date": "2025-01-15",
"comprehension_score": 85,
"reviews": [
{
"review_number": 1,
"scheduled_date": "2025-01-18",
"completed": true,
"completed_date": "2025-01-18",
"retention_score": 90
},
{
"review_number": 2,
"scheduled_date": "2025-01-29",
"completed": false,
"completed_date": null,
"retention_score": null
}
]
}
]
}
Review Session Workflow
When a review is due or user requests to review a chapter:
- Load chapter note file from
chapters/ - Present key questions from section 9 (Questions for Reflection)
- Add new questions based on cross-chapter connections discovered since initial reading
- Assess retention: Ask user to recall core teachings without looking at notes
- Score retention (0-100) and record in review-schedule.json
- Adjust future intervals: If retention < 70, schedule an additional review
Progress Tracking
Provide progress visibility when requested:
Overall Progress Command
When user asks "What's my progress?" or "How far have I gotten?":
## Reading Progress: [Book Title]
**Completion**: X/Y chapters (ZZ%)
**Chapters Completed**: 1, 2, 3, 5
**Current Chapter**: 6
**Chapters Remaining**: 7, 8, 9, 10, 11, 12
**Average Comprehension**: XX/100
**Total Reading Time**: XX hours
**Upcoming Reviews**:
- Chapter 1: Due in 2 days
- Chapter 3: Due in 5 days
**Current Streak**: Read X chapters in the last Y days
Chapter Status Overview
Provide a visual status overview:
## Chapter Status
1. Introduction ✓ (Score: 85)
2. Foundations ✓ (Score: 92)
3. Core Concepts ✓ (Score: 78)
4. Advanced Topics ⏸ (In progress)
5. Applications ○ (Not started)
...
State Management
To maintain continuity across sessions, track and persist:
Current Session State
Keep in working memory during active reading:
- Current book being read
- Current chapter number
- Current reading stage (pre/during/post)
- Output directory path
- Recently discussed concepts
Load Previous Session
When user returns to reading, check book-metadata.json to determine:
- Which book they're reading
- Last chapter accessed
- Reading progress
- Upcoming reviews
Greet returning users with context:
Welcome back! You're reading "[Book Title]".
**Last session**: Chapter X completed on [date]
**Current progress**: X/Y chapters (ZZ%)
**Suggested next**:
- Continue to Chapter Y
- Review Chapter A (scheduled for today)
File Update Protocol
After each meaningful interaction, update persistent files:
Always Update
book-metadata.json: Update timestamps, current chapter, status changes
Conditionally Update
- Chapter note files: When post-reading analysis is complete
glossary.md: When new terms are identified (batch updates preferable)cross-chapter-analysis.md: When completing a chapter that reveals connectionsreview-schedule.json: When completing a chapter or finishing a review
Write Strategy
- Use Write tool for new files
- Use Edit tool for updates to existing JSON/Markdown files
- Preserve all existing data when making updates
- Validate JSON syntax before writing
Usage Examples
Example 1: Starting a New Book
User: "I want to start reading 'Deep Learning' by Goodfellow. The PDF is at ~/Books/deep-learning.pdf"
Response:
- Ask for output directory
- Read PDF to extract TOC
- Create directory structure
- Initialize all metadata files
- Respond: "Initialized 'Deep Learning' with 15 chapters detected. Ready to begin. Which chapter would you like to start with?"
Example 2: Pre-Reading
User: "I'm about to read Chapter 3"
Response:
- Read Chapter 3 content
- Provide pre-reading brief with focus areas, key terms, roadmap
- Update book-metadata.json with current_chapter = 3
Example 3: During Reading Assistance
User: "I'm reading the section on backpropagation. Can you expand on the chain rule application?"
Response:
- Provide detailed explanation of chain rule in backpropagation context
- Offer analogies and examples
- Connect to previous chapters if relevant
- No file updates needed (just conversational assistance)
Example 4: Post-Reading
User: "I finished Chapter 3"
Response:
- Generate full 9-section chapter note
- Conduct comprehension testing (3-5 questions)
- Calculate score based on answers
- Save chapter note file
- Update book-metadata.json
- Add terms to glossary.md
- Update cross-chapter-analysis.md if applicable
- Schedule reviews in review-schedule.json
- Confirm completion and suggest next chapter
Example 5: Review Session
User: "What reviews are due?"
Response:
- Check review-schedule.json
- List chapters with reviews scheduled today or overdue
- Offer to conduct review session
- When user agrees, load chapter notes and quiz on key concepts
Resources
This skill uses references to maintain lean instructions while providing detailed specifications:
references/output-formats.md
Contains detailed templates and schemas for all persistent file formats, including:
- Complete JSON schemas for book-metadata.json and review-schedule.json
- Extended chapter note examples
- Glossary formatting guidelines
references/spaced-repetition-guide.md
Provides in-depth explanation of:
- Forgetting curve science
- Review interval calculations
- Comprehension scoring rubric
- Retention assessment techniques
Note: These reference files are loaded into context only when needed for detailed formatting questions or troubleshooting. The main SKILL.md provides sufficient guidance for typical usage.