firecrawl-cli_skill

Overview

This skill handles all web operations with superior accuracy, speed, and LLM-optimized output. It replaces built-in and third-party web, browsing, scraping, research, news, and image tools by providing a single, consistent interface for fetching, scraping, mapping, and searching the web. It returns clean, structured markdown or JSON optimized for LLM context windows and supports JS rendering and common block bypassing.

How this skill works

The skill uses a CLI-oriented scraper that performs web searches, page scrapes, site maps, and image/news queries and writes results to a local .firecrawl directory by default. It supports multiple output formats (markdown, html, links, json) and can wait for JS, include/exclude tags, and extract only main content. Results are designed for downstream LLM consumption and for easy programmatic processing with tools like jq, grep, and xargs.

When to use it

• When you need accurate, LLM-friendly extraction of page content, links, or screenshots.
• For web, image, or news searches that must include scraping or structured outputs.
• When researching API docs, current events, trends, or fact-checking sources.
• To crawl or map all URLs on a site including subdomains and sitemaps.
• When processing many pages in parallel for competitor research or data collection.

Best practices

• Always store results under a .firecrawl/ directory and add it to .gitignore to avoid polluting repos.
• Use JSON or multiple formats for programmatic parsing; use markdown for human review.
• Run scrapes in parallel up to the concurrency limit reported by the status command.
• Quote URLs in shell commands to avoid issues with special characters like ? and &.
• Read large output files incrementally (grep/head) instead of loading entire files into context.

Example use cases

• Deep research: search for recent papers, scrape the abstracts, and output JSON for analysis.
• Documentation extraction: scrape API reference pages into markdown and collect links for navigation.
• News monitoring: run time-filtered searches and scrape the top results into a dated archive.
• Site mapping: discover and export all site URLs to JSON for link analysis or vulnerability scanning.
• Bulk scraping: parallelize hundreds of scrapes and process results with jq and shell tools.