What problem does it solve? Reading public web pages that plain curl cannot render requires a scraping tool, but it is easy to misuse one by attempting logins, pasting API keys, or re-installing CLIs. This Skill defines a safe, authenticated, read-only path for turning one public URL or search query into dated markdown with a written receipt. ## Core Features & Use Cases - Authenticated status probe first: Runs firecrawl --status before any scrape and refuses to proceed if the CLI is unauthenticated, deferring re-login to a human. - Single read operation: Performs one firecrawl scrape (or search/map) per request, never crawling, interacting, or filling forms. - Receipt writing: Saves a dated receipt under /workspace/gb-proof/ quoting the URL, time, and a short excerpt for auditability. - Use Case: You need the markdown content of a public documentation page that curl cannot render. The Skill checks CLI authentication, scrapes the page once, and writes a receipt with the excerpt. ## Quick Start Use the firecrawl-read skill to scrape this public URL and save a dated markdown receipt under /workspace.
