Overview
book-to-skill solves a specific, familiar problem: you want an agent to work the way a particular book teaches, and dumping the PDF into context is both expensive and ineffective. Instead of feeding the whole text, it reads the source and produces a skill — a compact `SKILL.md` carrying the book's core mental models, decision rules and a chapter index, with the full chapters split into separate markdown files the agent loads only when it needs them.
The split is what makes it cheap. The README puts the entry file at roughly 4,000 tokens and each on-demand chapter at about 1,000, and claims a 24×–51× reduction against pasting the source into context. Alongside the chapters it writes a `glossary.md`, a `patterns.md` and a `cheatsheet.md`, so the agent has short reference surfaces for terminology and recurring patterns rather than having to re-read a chapter to recall one definition.
It accepts a path, a folder or a glob, so a book is the headline case but a documentation set or a pile of internal PDFs works the same way. Output targets the open Agent Skills standard, which means the result is usable from GitHub Copilot CLI, Amp, Claude Code and Hermes Agent rather than being locked to one client. The converter is Python and MIT licensed; optional per-format extractors (poppler for PDFs, ebooklib for EPUBs) are installed as needed. The MIT licence covers the converter, not whatever you feed it — check that you have the right to process a copyrighted book before you do.
What it does
- Converts PDFs, EPUBs, folders or globs into a structured agent skill rather than a raw text dump
- Two-tier output: a ~4,000-token `SKILL.md` with mental models plus ~1,000-token per-chapter files loaded on demand
- Generates `glossary.md`, `patterns.md` and `cheatsheet.md` as short reference surfaces
- Claims a 24×–51× token reduction versus pasting the source into context
- Emits the open Agent Skills format — works with Copilot CLI, Amp, Claude Code and Hermes Agent
- Installable through the skills CLI or by cloning straight into your skills directory
Getting started
Python 3 is required. Format-specific extractors (poppler for PDFs, ebooklib for EPUBs) are optional and installed per source type.
Install the skill
The skills CLI is the short path; cloning into your skills directory does the same thing manually.
npx skills add virgiliojr94/book-to-skill
# or manually
git clone https://github.com/virgiliojr94/book-to-skill.git ~/.claude/skills/book-to-skillConvert a source
Point it at a single file, a folder or a glob. The optional second argument names the skill it produces.
/book-to-skill <path|folder|glob> [skill-name]Check what it wrote
You get `SKILL.md` (mental models plus a chapter index), one markdown file per chapter, and `glossary.md`, `patterns.md` and `cheatsheet.md`. Skim `SKILL.md` first — it is what the agent reads every time.
Use the generated skill
The output follows the open Agent Skills standard, so install it into Claude Code, Copilot CLI, Amp or Hermes Agent the same way you would any other skill.
Commands and code are distilled from the project's own documentation — always check the official repo for the latest.
When to use it
- Give a coding agent a specific methodology from a book without pasting the book into every prompt
- Turn an internal handbook or a documentation set into a skill your team's agents can share
- Cut the token cost of reference-heavy work where only one chapter is relevant per task
- Build a small library of domain skills — one per book — instead of one sprawling system prompt
How book-to-skill compares
book-to-skill alongside other open-source agent skills & plugins tools AI/TLDR tracks, ranked by GitHub stars.
| Tool | Stars | What it does |
|---|---|---|
| Superpowers | ★ 297k | A composable skills plugin that installs a spec-first, TDD, subagent-driven development methodology into Claude Code, Codex, Cursor, Gemini CLI and other coding harnesses. |
| Skills for Real Engineers | ★ 282k | Matt Pocock's everyday agent skills for coding agents, covering alignment grilling, planning, code review and research — small, composable and meant to be edited. |
| ECC | ★ 276k | An installable plugin that adds 68 agents, 286 skills, hooks, rules, memory and an agent-config security scanner to Claude Code, Codex and other coding harnesses. |
| Karpathy Coding Guidelines | ★ 218k | A Claude Code plugin and CLAUDE.md rule set built from Andrej Karpathy's observations on LLM coding pitfalls: think before coding, keep it simple, make surgical changes, and work to verifiable success criteria. |
| Ponytail | ★ 159k | A skill and plugin that makes a coding agent take the laziest correct route — a seven-rung ladder from "does this need to exist?" to "the minimum that works" — while keeping validation, security and accessibility off the chopping block. |
| Caveman | ★ 111k | A skill plus local proxy that shortens an AI coding agent's prose while leaving code, commands and paths untouched — the rule file cuts what the agent writes, the proxy compresses what it reads. |
| Addy’s Agent Skills | ★ 104k | A pack of production engineering skills for AI coding agents, with nine lifecycle slash commands — /spec, /plan, /build, /test, /review, /ship — installable into 70+ agents. |
| book-to-skill | ★ 34.3k | Distil a technical book into an agent skill |