What problem does it solve? Web research and scraping tasks often involve unsafe URL handling, leaked API keys, unbounded crawls, and ad hoc transport switching. This Skill enforces a disciplined, CLI-first workflow for Firecrawl operations with strict URL validation, secret handling, consent gates, and evidence scrubbing. ## Core Features & Use Cases - Intent-based command routing: Maps each request to the narrowest Firecrawl CLI command (search, scrape, map, crawl, interact, parse, monitor, research, agent, doctor) across four intent paths covering live data, shipped integrations, repeatable deliverables, and recovery. - Central URL and argv safety gate: Enforces no-shell literal argv, HTTPS-only URLs, DNS resolution checks against private or reserved ranges, and per-redirect revalidation to prevent SSRF. - Secret and evidence hygiene: Prohibits API keys in argv, mandates secure credential stores, and requires scrubbed, permission-restricted evidence files with retention and cleanup tracking. - Use Case: A user asks to collect every page in a site's /guides section for a market brief. The Skill selects the crawl path, validates all URLs through the safety gate, bounds the crawl scope, and retains only scrubbed source URLs and assertions as traceable evidence. ## Quick Start Ask the agent to scrape a known public URL with Firecrawl, and it will verify the installed CLI, apply the URL safety gate, and run the narrowest scrape command.
