The skill configures stealth mode to avoid bot detection, setting appropriate user-agent strings, viewport dimensions, and JavaScript execution timing. It handles wait conditions — waiting for specific selectors, network idle states, or custom JavaScript expressions before extracting content.

Connection management is built in: the skill tracks concurrent connections against the plan limit, queues excess requests, and manages session tokens for authenticated scraping flows. It supports cookie injection for accessing content behind login walls.

For complex extraction, the skill injects custom JavaScript functions that run in the page context to extract structured data from DOM elements. Results are returned as clean JSON objects. The skill also supports generating PDFs of rendered pages for archival purposes. Rate limiting and retry logic handle transient failures gracefully. Best suited for scraping modern web applications that resist traditional HTTP-based extraction.