Claude Command (Imported): crawl

• Source: C:\Users\monke\.claude\plugins\cache\claude-plugins-official\firecrawl\1.0.0\commands\crawl.md
• Plugin: firecrawl
• Version: 1.0.0

Notes

• Claude command files may use ! prefixes for shell execution. In Codex, run those commands in the terminal manually.

Original Command Frontmatter (Reference)

description: Crawl an entire website and extract content from multiple pages
allowed-tools: firecrawl_crawl, firecrawl_check_crawl_status

Original Command Body

Crawl Website

Use the Firecrawl crawl tool to extract content from multiple pages of a website.

Important: Auto-poll for completion

Crawl operations are asynchronous. When you start a crawl:

1. You'll receive an operation ID
2. Automatically call firecrawl_check_crawl_status with that ID
3. Keep checking until status is "completed" or "failed"
4. Return the final results to the user

Do NOT ask the user to check status manually - handle it automatically.

When crawling:

1. Ask the user for the starting URL if not provided
2. Clarify the scope: how many pages, specific sections, or entire site
3. Use the firecrawl_crawl tool with appropriate parameters
4. Poll for completion using firecrawl_check_crawl_status
5. Summarize the crawled content and provide structured results

Available Formats

Same as scrape - you can request:

• markdown, html, rawHtml, screenshot, links, summary

Options

• maxDepth: How deep to crawl (default: 2)
• limit: Maximum number of pages to crawl
• allowExternalLinks: Whether to follow external links
• includePaths: URL patterns to include
• excludePaths: URL patterns to exclude
• deduplicateSimilarURLs: Skip similar URLs

Tips

• Consider using the /map command first to discover available URLs
• Set reasonable limits to avoid long wait times
• Use includePaths to focus on specific sections