book-to-skill

book-to-skill is an open-source Agent Skill by virgiliojr94 that turns a technical book, a document folder or a set of papers you own into a structured skill your agent can study and reference while you work. Instead of searching a PDF or hoping the agent remembers the book, you ask the generated skill about a topic and it reads the right chapter notes. What you can do with it - Convert a book: /book-to-skill ./my-book.pdf creates a skill with the book's core mental models, a chapter index, per-chapter notes, a glossary, a patterns file and a cheat sheet - Ask the result about a topic or chapter, such as /designing-data-intensive-apps replication - Merge several files, a folder or a glob into one skill, and fold new material into an existing skill later - Use it beyond books: internal docs, runbooks, brand and design guidelines, specs and research papers - Works with PDF, EPUB, DOCX, HTML, Markdown, plain text, RTF and, with Calibre, MOBI and AZW How it works - A Python extractor turns the document into clean text on your machine; your agent then follows the skill's steps to write the new skill. - Chapter files load only when you ask about that topic, which keeps the loaded skill small. - Install it with npx skills add virgiliojr94/book-to-skill, or clone it into your agent's skills folder. It follows the open Agent Skills standard; the maker documents Claude Code, Codex, GitHub Copilot CLI and more. - Optional extractors (Docling for technical books with code and tables, pdftotext, ebooklib) improve quality. Scanned PDFs need OCR first. Good to know - It ships no book content. You point it at files you have the right to use, and the generated skill is synthesized notes: the skill's own rules say never to copy raw book text. - The maker asks you to keep skills made from third-party copyrighted books private. If you choose to publish a generated skill to GitHub, the repository is private by default. - Extraction runs locally; the text your agent reads goes to its model provider like any other prompt. - MIT licensed (the converter, not the books you process). A community project by virgiliojr94.

• “/book-to-skill ~/books/designing-data-intensive-applications.pdf”

  Copy

• “Turn our docs/ folder of runbooks and ADRs into one skill called project-knowledge”

  Copy

• “Merge these three papers and my notes into a unified research skill”

  Copy

• “Fold this new paper into my existing project-knowledge skill”

  Copy

• “/designing-data-intensive-apps replication”

  Copy

• “Which chapters does my DDIA skill cover?”

  Copy