structured-data-extractionlisted

Extract data from the machine-readable formats sites already publish — JSON-LD, microdata, RDFa, Open Graph, RSS, Atom, sitemaps and hydration blobs — before writing any selector. Use whenever the user is scraping a page, wants product or article or event data, mentions JSON-LD or schema.org or microdata or RSS or feeds, or asks for the most reliable way to get data off a site.
manypicom/web-data-skills · ★ 0 · Data & Documents · score 72

Install: claude install-skill manypicom/web-data-skills

# Structured Data Extraction A large share of the web already publishes the data you want in a typed, machine-readable format, put there deliberately so that machines can read it. Most scrapers ignore it and write CSS selectors against the rendered markup instead. **Check the structured formats first, every time.** It's faster to write, far more stable, semantically labelled, and it usually gives you fields the visible page doesn't show. ## The formats, in order of usefulness | Format | Where | What you get | |---|---|---| | **JSON-LD** | `<script type="application/ld+json">` | Typed entities. Often the complete record | | **Hydration blob** | `__NEXT_DATA__`, `window.__NUXT__`, inline JSON | The page's own data model. Richest of all | | **Microdata** | `itemscope` / `itemprop` attributes | Typed, inline, older but common | | **RDFa** | `vocab` / `typeof` / `property` | Same idea, less common | | **Open Graph / Twitter** | `<meta property="og:*">` | Title, description, image, type | | **RSS / Atom** | `/feed`, `/rss.xml`, `<link rel="alternate">` | Clean list of items with dates | | **Sitemaps** | `/sitemap.xml`, `robots.txt` | The site's own URL inventory | | **JSON API** | Network requests behind the page | Paginated, typed, most stable of all | JSON-LD is the highest-value routine check. Hydration blobs and the underlying API are higher value still when present, and both are covered in `javascript-rendered-scraping`. ## Finding what a page publishes ```bash u="https