The author’s Booklin wizard banner from the book-to-skill README; project artwork, rather than a generated book Skill.
Project screenshotsAbout this Skill
Book to Skill combines a Python document extractor with a root SKILL.md that guides your AI agent through knowledge organization. It accepts a file, directory, glob or list of sources and can produce one unified Skill with named frameworks, principles, techniques and anti-patterns. It also supports analysis-only, generation from existing analysis and folding new material into an existing Skill.
Why we picked it
Useful when you repeatedly consult the same material while coding, studying or writing. A local Python extractor prepares the text, then your agent organizes frameworks, decision rules and references into a Skill that can load relevant chapters as needed. The full conversion uses an AI agent; installing the standalone extraction CLI alone does not register the Agent Skill.
What it helps with
- Readers who want to consult and apply the frameworks of a specific book while working.
- Teams organizing their own runbooks, architecture decisions, onboarding documents or design guidance.
- Researchers combining a related set of papers and notes into an evolving reference Skill.
From source documents to a reusable Skill
The extractor reads local documents into text and metadata. Your agent then analyzes the structure and writes the Skill files. Choose text mode for mostly prose or technical mode when tables, code and formulas need structure-aware extraction. This is a guided agent workflow rather than a standalone PDF-to-Skill command that performs the entire generation by itself.
Flexible inputs
PDF, EPUB, DOCX, TXT, Markdown, reStructuredText, AsciiDoc, HTML, RTF and Calibre-supported MOBI/AZW formats; files, directories, globs and lists can be combined.
Conversion modes
Run a full conversion, stop after analysis, generate from prior analysis, or fold new sources into an existing Skill.
Agent Skill entry pointUsage, modes and update examplesConversion workflow
Files the agent can generate
The output structure is designed for topic-based access: the main SKILL.md carries core models and a chapter index, while chapter notes and reference files remain available for targeted loading. The actual depth depends on the source and your agent’s conversion.
SKILL.md and chapters/
A discoverable entry point with mental models and a topic index, plus chapter files describing the relevant frameworks, examples and techniques.
glossary.md, patterns.md and cheatsheet.md
Terms with chapter references, reusable techniques and patterns, and decision tables or quick-reference rules.
Install the Agent Skill
| Command | What it helps with |
|---|---|
git clone https://github.com/virgiliojr94/book-to-skill.git .claude/skills/book-to-skill | Claude Code project install. Entry: .claude/skills/book-to-skill/SKILL.md. |
git clone https://github.com/virgiliojr94/book-to-skill.git .agents/skills/book-to-skill | Codex project install. Entry: .agents/skills/book-to-skill/SKILL.md. |
Install the complete repository
Run the skills CLI and select your host. Preserve the extractor modules and tools with SKILL.md; copying only the entry-point Markdown omits runtime files.
npx skills add virgiliojr94/book-to-skillRefresh discovery
Start a new agent session after installation. In GitHub Copilot CLI, the upstream guide also provides /skills reload and /skills info book-to-skill.
Convert a document
Pass your document path and an optional output slug to the Skill. Generated book Skills use the converter’s personal destination by default; project-local output is an explicit choice.
/book-to-skill ./my-book.pdf my-book-notes
Start with a file, folder or update
You can also ask for analysis only when you want to review the extracted frameworks before creating any Skill files. Check the generated chapter names, terminology and examples against the original material.
| Command | What it helps with |
|---|---|
/book-to-skill ./my-book.pdf my-book-notes | Convert one document into a named Skill. |
/book-to-skill ./project-docs/ project-knowledge | Combine supported files from a documentation folder. |
/book-to-skill ./new-notes.md /path/to/existing-skill | Fold new material into an existing Skill folder. |
Optional extraction CLI and dependencies
The standalone CLI installs the extraction engine, not the Agent Skill. The installation guide obtains it directly from GitHub; the skills CLI or complete repository installation remains the route for the agent-driven conversion. From a repository checkout, scripts/extract.py --check reports available extractors.
| Command | What it helps with |
|---|---|
python3 scripts/extract.py --check | Check format-specific tools from the repository checkout. |
pip install "book-to-skill[pdf,epub,docx] @ git+https://github.com/virgiliojr94/book-to-skill.git" | Install the optional extraction-only CLI from GitHub. |
book-to-skill ./my-book.pdf --mode text | Extract text with the standalone CLI; this does not generate a complete Agent Skill by itself. |
Official repository READMEInstallation and host guidePython package and optional dependencies
Extraction scope, source rights and project artwork
Scanned PDFs need a text layer from a separate OCR step before conversion. The upstream FAQ also states that text embedded in diagrams or charts is not extracted. Text extraction runs locally; when a cloud-hosted model analyzes extracted text, that input follows your AI provider’s data terms.
The repository’s MIT license covers the converter code and Skill definition. It does not relicense the books or documents you process. Source documents and generated material remain subject to their applicable terms.
The gallery reuses the author’s Booklin banner from the official README. It is project artwork and represents the converter; it is not a screenshot of a generated book Skill. The supplied WebP is already below the catalog’s static-image budget.
Official repository READMEOfficial FAQMIT licenseAuthor README banner
Before you use it
- A compatible coding agent with shell execution and local file access, plus Python 3.9 or newer according to the package metadata.
- Install format-specific dependencies as needed. Technical PDF extraction uses Docling; text PDFs can use pdftotext, pypdf or pdfminer.six. MOBI/AZW/AZW3 require Calibre.
- Use source documents you are entitled to process. The converter is MIT-licensed and available without purchase; AI services and source documents have their own costs and terms.
Supported agents
Frequently asked questions
Does installing the pip CLI create /book-to-skill in my agent?
No. The pip package provides the extraction engine. Install the complete Agent Skill with the skills CLI or a full repository clone to get the agent-driven conversion workflow.
Can it merge multiple sources or update an existing Skill?
Yes. The documented workflow accepts multiple paths, a folder or a glob, and has a separate update/fold-in mode for adding material to an existing Skill.
Does it process scanned books and chart images directly?
Scanned PDFs require OCR beforehand. The converter focuses on text extraction and organization; the upstream FAQ says text inside diagrams and charts is not extracted.
Official sources
Based on official documentation, with practical scenarios suggested by SkillHey.
