wtf is book-to-skill?

book-to-skill turns a technical book, or really any structured document such as a PDF, EPUB, DOCX, or even a folder of files, into a structured skill that an AI coding agent like Claude Code, GitHub Copilot CLI, or Amp can load and use while you work. Instead of pasting a whole book into the agent's context, or hoping the agent remembers something it half read once, you convert the book once, and the agent then loads only the specific chapter it needs, on demand, when you ask it a question. The idea it responds to is a common one: you read a great technical book once, and months later you cannot remember what a particular chapter covered. Searching the PDF gives you a list of pages, not an answer, and asking an agent about a book it never saw tends to produce guesses instead of real content. This tool solves that by running the book through an extraction step that produces a skill folder containing a core summary file, one file per chapter, a glossary of key terms, a file of patterns and techniques, and a quick reference cheat sheet. Chapter files only get loaded when you actually ask about that topic, so the ongoing cost stays small even though the whole book was processed once. The README reports that this approach uses 24 to 51 times fewer tokens than simply dumping the full book into context to answer a single question, based on measurements against real books. Beyond books, the same process works on internal documentation, brand or design guideline documents, collections of research papers, or specifications and standards, basically anything you reference often enough that you wish you had memorized it. Getting started is a single command, either through a cross agent skills installer or by cloning the repository directly into your agent's skills folder. The project is written in Python and is released under the MIT license, so it is free to use, modify, and reuse for any purpose.