firecrawl-crawl: Claude Code Skill & Agent Tool
An open-source skill by firecrawl (opens in a new tab), published in the firecrawl/cli (opens in a new tab) repository on GitHub. Bulk-extracts many pages from one site or section — for example everything under /docs — with depth limits and status checks for long-running crawls.
Install firecrawl-crawl
Requires Node.jsnpx skills add firecrawl/cli --skill firecrawl-crawlRun this command in your project terminal to install this skill for Claude Code.
What this skill does
firecrawl-crawl is the Firecrawl CLI skill for bulk extraction. When one page is not enough — you need everything under /docs, a full blog archive, or content spread across linked pages — your agent starts a crawl that follows links from a starting URL up to a depth and page limit you control, and saves the result as JSON in a .firecrawl folder.
The skill teaches the agent to scope every crawl. When your request names a section, it uses include-paths so only the pages you need are fetched; for a whole site, it sets a maximum depth and a page limit. It can wait for the crawl to finish with a progress display, bound that wait with a timeout, or run asynchronously and check the job's status later by its ID. The job counts as done only when the crawl reaches a terminal status and the saved output contains the expected pages.
Crawling consumes Firecrawl credits per page, so the agent checks your credit usage before large jobs, and crawls require a signed-in Firecrawl account. Use it with Claude Code to build a knowledge base from documentation or to migrate a site's content.
Key Capabilities
- Crawls from a start URL, following links up to a depth and page limit
- Scopes crawls to a section such as /docs with include-paths
- Waits with progress and an optional timeout, or runs asynchronously by job ID
- Saves results as JSON under .firecrawl/ and checks the expected pages are there
- Checks your credit usage before large crawls
How to use with Claude Code & AI Agents
Install the skill
Run the command above at the root of your project. The CLI asks which agents to set up — Claude Code, Cursor and other supported agents — and adds firecrawl-crawl to the project, so teammates who pull the repository get it too. Add-gto install it for your user account across every project.Open a new agent session
Start a fresh Claude Code session in that project so the agent loads the new skill. Have ready: Firecrawl CLI; Signed-in Firecrawl account (crawl has no keyless tier).Ask in plain words
Describe the task, for example:“Crawl everything under https://docs.example.com/docs and save it as JSON.”The agent matches your request to the skill's description and follows its SKILL.md instructions step by step. Read them on GitHub before using the skill on a system that matters.