Ten self-contained Node.js Playwright scripts, one file per scraping pattern: paginated tables to CSV, infinite scroll feeds, login-then-scrape with saved sessions, price and stock monitors with webhooks, bulk form submit, job boards with cross-run dedupe, a robots-aware sitemap crawler, screenshot diffing, JSON polling, and a retry wrapper with backoff.

Each script runs on its own: how-to comment on top, selector constants marked EDIT, --flag arguments, progress printed. Copy a file, change the selectors, run it. Needs Node 18+ and one chromium install; no API keys and no accounts (the login template reads credentials from env vars). A smoke-test runs four scripts on a local fixture site.

Fair use is built in: delays of 800 to 2500 ms in every script, the crawler honors robots.txt and adopts a stricter crawl-delay, no captcha solving, no credential stuffing, no IP-rotation evasion.