reference-harvester

@drpaulcoleman

Download web pages, PDFs, and search results into the local references/ folder for AI context. Runtime-adaptive — detects and uses whatever is installed (Python+Playwright, Node.js+Playwright/Puppeteer, PowerShell+Edge, or plain curl). Use this skill whenever the user wants to fetch a URL, capture documentation, save an article or PDF, gather sources on a subject or topic, "download" reference material, JS-render a page that came back as an empty shell, or runs the /download command. Prefers a linked PDF over crawling many HTML pages; always includes Google Scholar when searching a subject.