ultra-fetchlisted

Research a topic from web sources and secure the originals — fetch, crawl or map pages into clean, context-efficient markdown saved to a file, using a stealth browser that reaches sites the built-in WebFetch can't (bot-protected, Cloudflare, JS-rendered). This is the default way to read a web page's content, starting with a single page whose URL you already have, and extending to opening what a WebSearch turned up, reading across a whole site, or discovering what URLs a site has. Also the recovery when WebFetch hands back a refusal instead of a page — it reports a block as ordinary body text on a call it marks successful, and Korean news sites (yna.co.kr, mk.co.kr, fnnews.com) refuse it outright. NOT for a trivial quick fact where WebFetch already suffices, NOT for anything behind a login — it carries no credentials and no session, so authenticated pages, private feeds and paywalled articles are out of reach — and NOT for developing the ultra-fetch tool itself, which is ordinary repo work.
tjdwls101010/Ultra-Fetch · ★ 0 · Data & Documents · score 70

Install: claude install-skill tjdwls101010/Ultra-Fetch

# Ultra Fetch ## The division of labor `ultra-fetch` gets you **access** (a stealth browser that reaches what WebFetch can't) and **clean output** (markdown with the boilerplate stripped, written to a file). Everything else is your judgment: which command fits the question, how deep to go, how much of the result to actually read. Reach for it by default when you need a page's real content. Stay away from it for a fact you already know, for anything behind a login, and when the task is editing this tool's own code. ## Starting from a search The usual shape of a research task is WebSearch first, this tool second: search returns leads, and a lead becomes a source only once you have opened it. Two things about that hand-off will cost you if you meet them cold. **WebFetch reports a block as though it were a page.** The tool call is marked *successful* and the refusal arrives as ordinary body text — "The server returned HTTP 403 Forbidden", or "domains not accessible to our user agent". Nothing about it is shaped like an error, which is exactly how a source gets dropped mid-sweep without anyone noticing. Its own advice points away from the fix, too: it suggests an "authenticated tool", but these sites are not asking who you are, they are refusing a client that isn't a browser. A stealth browser is what gets through, and that is this tool. Worth remembering by name, because they were measured and they recur: **`yna.co.kr`, `mk.co.kr` and `fnnews.com` refuse WebFetch.** Korean