research / skill

book-to-skill

An extractor and agent skill for converting technical books, document folders or collections of sources into skills with chapter files, a glossary, patterns and a cheatsheet.

Repository owner virgiliojr94

Not rated yet
0 comments

Like 0 Save

Published 11 Oct 2026Listing updated 12 Oct 2026

At a glance

Purpose: An extractor and agent skill for converting technical books, document folders or collections of sources into skills with chapter files, a glossary, patterns and a cheatsheet.
Runs in: Local agent with shell
Operating systems: Not stated
Needs: Runtime: Python; Needs internet: Not stated; Needs accounts or API keys beyond your agent: Not stated
Files and data: Reads files: Yes; Writes files: Yes; Sends data to other services: Yes
Tool type: Skill
Delivery: From GitHub
Version: Latest on the master branch
Access: Code on GitHub
Compatibility: According to the README: GitHub Copilot CLI, Amp, Claude Code, Hermes Agent, OpenCode and OpenClaw, and any host that supports the open Agent Skills standard. It also names Codex among the tools that discover the shared skills folder. The README describes different install locations by host.
External costs: The repository is MIT-licensed. The README says book-to-skill is free and maintained on personal time, and asks for optional sponsorship. Your coding agent or model provider may charge for usage.
Permissions: The skill acts through your coding agent with the access you give it. The README describes downloading the skill through npx skills add or cloning it into a user-level agent skills folder. The README says conversion reads selected files or folders and writes generated skills into ~/.agents/skills/<slug>/ or a host-specific destination, with an attempted symlink under ~/.claude/skills/<slug>/ for Claude Code. The README lists other home-directory locations for Copilot CLI, Hermes, OpenClaw and OpenCode. The README shows apt and pip3 commands for format-specific extractors and identifies Calibre as an external application. The README says a generated skill can be published to GitHub, private by default.
Usage terms: MIT License, as stated in the repository. See its LICENSE.md file for the exact terms. The README says the licence applies to the converter in this repository, not to any book or document you process with it.

Code on GitHub

virgiliojr94/book-to-skill

MIT

Python34,666 stars3,637 forksLast pushed 12 Oct 2026 UTC

Turn any technical book PDF into a Claude Code skill — ready to study, reference, and use while you work.

Topics: agent-skills ai-agents book-to-skill context-engineering document-processing edtech knowledge-base knowledge-management

Public GitHub data fetched 13 Oct 2026 UTC · 0 days old

View on GitHub ↗

What it does

book-to-skill is a converter for technical books, document folders and collections of sources. The README describes a Python extractor and a generator whose instructions your agent follows. The README gives /book-to-skill with a file, folder or glob as its invocation and says the generated skill contains frameworks, decision rules, anti-patterns and per-chapter files. The README says an agent can load the relevant chapter on demand.

The README lists generated files: SKILL.md with core mental models and a chapter index, chapter files, glossary.md, patterns.md and cheatsheet.md. The README says chapter files do not count against the skill budget until requested. The README describes PDF, EPUB, DOCX, Markdown, HTML, RTF and MOBI inputs, along with internal documentation, brand and design systems, research papers and notes, and specs and standards.

The README says extraction and analysis run on your machine and that users should bring material they have the right to process. The README warns that publishing or sharing a generated skill from a copyrighted work can infringe the rights holder and says to keep skills from third-party books private.

Who it is for

People who want to study or look things up in a technical book, internal docs or specs from inside a coding agent.

Example tasks

Convert a PDF book into a skill with /book-to-skill ./my-book.pdf
Ask a generated skill about a topic, as in /my-book replication
Fold a docs folder of runbooks or architecture decision records into one skill
Check which extractors are installed with python3 scripts/extract.py --check
Run ocrmypdf input.pdf output.pdf on a scanned PDF before converting it

Getting started

The README gives npx skills add virgiliojr94/book-to-skill as the cross-agent installation command, or cloning the repository into your agent's skills folder as an alternative. The README lists folders for Claude Code, Copilot CLI, Amp, Hermes Agent, OpenClaw and OpenCode. The README says python3 scripts/extract.py --check lists installed extractors by format and commands to install missing tools.

Limitations

The README names GitHub Copilot CLI, Amp, Claude Code, Hermes Agent, OpenCode and OpenClaw, and describes hosts supporting the Agent Skills standard. The README says Hermes uses category-specific personal skill folders rather than the cross-agent root. The README says OpenClaw discovers ~/.agents/skills/ with its default state directory; a non-default OPENCLAW_STATE_DIR requires that state's skills root. The README says Claude Code discovery depends on successful readback of the attempted symlink. The README describes a Python extractor with format-specific tools and fallbacks, and Calibre's external ebook-convert app for MOBI, AZW and AZW3. The README says scanned PDFs need OCR first and that the skill asks whether a book is technical or text-heavy before extraction. The repository can change at any time; check its README before you install.

Data practices

The README says extraction and analysis run on your machine. The README adds that when your agent's model runs in the cloud, text supplied to it follows that provider's normal data terms. The README says the converter can publish a generated skill to GitHub, private by default, and warns against redistributing skills generated from copyrighted third-party books.