smerchek / claude-epub-skill

The one skill, markdown-to-epub-converter, treats H1 headers as chapter breaks and H2 through H6 as sections, carries over bold, italic, links, lists, code blocks, and tables, and builds both an NCX and a NAV table of contents so the result opens correctly across Apple Books, Kindle apps, Kobo, and other EPUB3 readers. It also reads a YAML frontmatter block for title, author, language, and date, so a document that already has that metadata keeps it in the finished ebook. Two Python scripts do the actual work — markdown_processor.py for parsing and epub_generator.py for building the file with the ebooklib library — and the repo's own test suite covers empty content, plain text, and documents over ten chapters long.

This is for someone using Claude, or Claude Code, who wants a written markdown document, a set of research notes, or a chat summary turned into a portable ebook rather than left as a text file. It needs Python 3.8+ and two pip packages (ebooklib and markdown2) already on the machine; nothing here calls out to a paid service or asks for a key.

READ THE FULL ANALYSIS＋

What the README lists but the skill doesn't yet do. Its own "Future Enhancements" section names a cover-page generator, native Kindle formats (.mobi and .azw3), custom fonts, and merging multiple markdown files into one book — none of that exists today. What ships is exactly the plain markdown-to-EPUB3 conversion described above, nothing more.

Checked 19 September 2026 from the repository's own README and its single skill description; we did not run the conversion ourselves.