scrape

Web scraping and browser automation with Playwright — always browser-based, never fetch/axios

RULE: This command ALWAYS uses Playwright. Never use fetch, axios, requests, cheerio, or puppeteer.

Input: $ARGUMENTS — URL(s) and/or task description.

Apply the dog:playwright-scraping skill to implement browser-based scraping or automation.

The skill will:

  1. Analyze the task: target URL(s), data to extract, auth requirements, pagination, output format
  2. Generate TypeScript + Bun scraper using Playwright chromium
  3. Handle: JS rendering, authentication flows, dynamic content, pagination, file downloads
  4. Block unnecessary resources (images, fonts) for performance
  5. Structure output as JSON by default; CSV or markdown on request
  6. Handle errors gracefully: retry on timeout, screenshot on failure

For complex automation flows (login, form submission, multi-step): describe the full flow in $ARGUMENTS.