HiAi-gg Scrape a Page with Firecrawl
Use this skill when an agent needs the clean content of a specific URL.
Workflow
1. Confirm the target
- Get the exact URL from the user or from a search result.
- State what you expect to extract (full text, specific section, structured fields) so you can verify the output.
2. Scrape
firecrawl_scrape(url)
The result is normalized, agent-readable content (markdown by default).
3. Verify the extraction
- Check the content actually covers the target page (correct page, no captcha wall, no empty body).
- If the page is interactive-heavy (JS-rendered SPA), note that the scrape may miss content; prefer a static URL or use browser interaction tools where available.
- If you need only part of the page, extract that part rather than dumping the whole body into the response.
4. Report
Give: the source URL, a summary of the content (or the requested excerpt), and any caveats (paywall, partial render, robots-limited). Do not include the full page text unless the user asked for it.
Guardrails
- Scrape only URLs the user provided or that came from the user's task.
- Respect site terms and robots directives; do not bypass paywalls or access controls.
- Never pass credentials through the scrape call.