If you're tired of fighting with PDFs and Word docs to extract clean, structured data, this is worth a look. Docling handles PDFs, PowerPoint, images, and more with what they call "advanced layout understanding," which in practice means it actually preserves document structure instead of dumping garbage text at you. The quick start is genuinely quick: point it at a URL or file path, call convert, export to markdown. It's from the existential-birds/beagle collection and has decent traction with 434 installs. Passes most security audits except a Snyk warning you'll want to check. Good fit if you're building anything that ingests documents and needs the content in a usable format without manual cleanup.
npx skills add https://github.com/existential-birds/beagle --skill docling