AI Pulse · Under the Radar

A small tool called book-to-skill popped up on GitHub today and it’s already the second fastest growing repo in the whole platform, picking up around 859 stars a day. That kind of pace usually means developers are hitting a real problem, not just chasing something shiny.

Here’s what it does. Feed it any technical book, PDF, EPUB, DOCX, HTML, RTF, or MOBI, and it restructures the content into what’s called a Claude Code or Copilot CLI “skill.” Instead of dumping an entire 300 page manual into an AI’s context window every time you need one answer, the tool breaks the book into a core mental model file, plus chapters, a glossary, common patterns, and a cheatsheet you can pull up on demand. The AI reads what it needs, when it needs it, rather than chewing through the whole thing every single time.

For a business owner, the direct use case is obvious if you’ve built any training material, SOPs, or course content. If you or your team have written booklets, internal manuals, or course workbooks, this is a way to turn that content into something an AI coding assistant or internal tool can actually reference properly, instead of you copying and pasting sections into a chat window every time someone has a question. It’s the difference between a library with a card catalog and a library where you have to read every book cover to cover to find one paragraph.

This is early and unproven at scale, but it points at a bigger shift. Businesses are going to want their own documentation turned into structured, queryable knowledge rather than static PDFs sitting in a shared drive. That’s exactly the kind of thing we build into an AI command centre, where your internal knowledge becomes something your tools can actually use, not just store.