- Home
- Skills
- Productivity
- book buddy
book buddy
A spoiler-safe reading companion that indexes local books and discusses them only up to your current page.
Free
Works with the AI Agents you already use
book buddy
Example session with this skill installed
I am on page 142 of 'Project Hail Mary'. Who is Ryland Grace talking to right now, and what have we learned about his past so far? Cite the pages.
- Read your context and instructions
- Compiled the book buddy
I've searched the indexed text up to page 142. Here is a summary of Ryland's current situation and his background as revealed so far.
Reading Brief: Project Hail Mary (up to p. 142)
| Topic | Details | Source |
|---|---|---|
| Current Interaction | Ryland is currently communicating with an alien entity he has nicknamed "Rocky." They are using a system of musical notes and frequencies to establish a basic vocabulary. | (p. 138) |
| Early Career | Before being recruited, Grace was a junior high school science teacher. He left professional academia after a controversial paper regarding the necessity of water for life. | (p. 42) |
| Recruitment | He was forcibly recruited into the Petrova Task Force by Eva Stratt because of his specific expertise in molecular biology and his "out of the box" theories. | (p. 68) |
| Memory Recovery | Grace initially woke up with total amnesia; his name and profession were recovered through a series of physics experiments he performed in the lab. | (p. 12-15) |
Next steps
- Update your progress with
python3 scripts/bookbuddy.py progress --set [N]when you reach the next chapter. - Ask for a "Character Tracker" if you want a full timeline of Ryland's interactions with Stratt.
- Use the
readcommand to get a summary of the specific "flashback" sequence from pages 60-75.
Connects securely to your tools. The creator never sees your data.
About this skill
Stop worrying about spoilers when discussing books with an LLM. Most models are trained on literature and will accidentally reveal plot twists or endings from later chapters. This skill enforces a strict reading boundary by indexing your local book files and only allowing the agent to "see" and discuss the pages you have already finished.
What it does
- Spoiler protection prevents the agent from discussing any plot points, characters, or events beyond your current page or location.
- Local indexing processes EPUB, PDF, MOBI, and TXT files locally using a SQLite index with zero external network calls.
- Page-level citations requires the agent to ground every claim with a specific page number or location ID from the text.
- Progress tracking maintains your current position in the book and adjusts the search window as you read further.
- Character tracking traces character mentions and motivations exclusively within the context of what you have read so far.
How it works
- Ingest the file using the
ingestcommand to create a local index of the book's text and page structure. - Set your progress by providing your current page number or approximate location to lock the spoiler-safe boundary.
- Ask questions while the agent executes multiple keyword searches (BM25) to retrieve relevant passages.
- Summarize or recap by reading specific ranges of text to catch up on recent events without looking ahead.
Frameworks & tools
This skill uses Python 3.9+ with pypdf for PDF parsing and mobi for MOBI file support. Data is stored in a local SQLite index.
Why this beats prompting it yourself
Standard prompts cannot stop a model from using its pre-trained knowledge of famous plot twists. This skill solves that by physically restricting the context window to only provided passages and forcing the agent to ignore its internal memory in favor of the indexed file.
Use cases
- Thematic analysis without discovering which characters die in later chapters.
- Character refreshers to remember who a minor person is without seeing their future arc.
- Reading recaps for people returning to a book after a long break.
- Study assistance requiring precise page citations for academic or book club preparation.
Known limitations
Scanned PDFs without a text layer are not supported and require OCR first. EPUB and TXT page numbers are approximate locations based on character counts.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Productivity.
Scanned PDFs without a text layer require external OCR. EPUB, MOBI, and TXT use approximate locations (2,000 characters) rather than physical page numbers.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed today
- Free with an Agensi account
- Software
Python