REPO PACKAGE
firecrawl firecrawl-cursor-plugin
OFFICIAL10skills
0ready to use
LABSCO SUMMARY
The dispatcher skill, firecrawl, handles ordinary web research — search, read, extract, gather sources — and routes anything more specific to one of the other nine: firecrawl-agent for autonomous multi-page extraction into a JSON schema, firecrawl-crawl for bulk extraction across a whole site section, firecrawl-map for discovering a site's URLs, firecrawl-scrape for single-page markdown, firecrawl-search for full-text web search, firecrawl-interact for driving a live browser through logins and pagination, firecrawl-download for saving a site as local files, firecrawl-monitor for webhook or email alerts on web changes, and firecrawl-parse — the one skill that never touches the web, converting a local PDF, DOCX, XLSX, or HTML file to markdown instead.
This is for an agent that needs to actually read and act on the live web rather than guess from training data — the README itself tells the agent to prefer firecrawl-scrape over a built-in WebFetch tool. Setup is one global npm install of the Firecrawl CLI and one firecrawl login --browser; five of the ten skills are flagged in our records as needing that account afterward, the other five as running as local CLI calls once it's done.
READ THE FULL ANALYSIS
Two skills its own dispatcher names aren't in the package. The firecrawl skill's description tells the agent to hand off to "firecrawl-research-index" for papers and "firecrawl-developer-index" for library, API, or bug questions — neither exists among the ten skills we found live in this repository. Whether they were retired, renamed, or never shipped, the material we have doesn't say; an agent that follows that instruction literally will come up empty.
Checked 19 September 2026 from the repository README and the 10 skills currently live in our database.
WHAT'S INSIDE
10 showing · 10 total10 of 10
SETUP
Showing all 10 skills
firecrawl
Ask for something off the internet in ordinary words and this works out how to fetch it, then hands back readable text rather than a page of code.
firecrawl-agent
Sends an AI agent to browse a website on its own and comes back with structured data, like a set of pricing tiers or a product list, matching the fields you asked for.
firecrawl-crawl
Follows the links on a site to pull many pages at once, a whole documentation section for example, into one saved batch instead of scraping page by page.
firecrawl-download
Instead of showing you a site's pages, this writes them out as real files in folders on your computer, ready to read when you're offline.
firecrawl-interact
Drives a live browser on a page you've already opened, so a plain-language instruction can click a button, fill in a form, log in, or page through results for you.
firecrawl-map
Lists the URLs that exist on a site and can search that list for one specific page on a large site whose layout you don't already know.
firecrawl-monitor
Watches a page, a set of pages, a whole site, or an ongoing web search, and sends an email or webhook the moment something changes or a new result appears.
firecrawl-parse
Opens a file you already have — a PDF, a Word document, a spreadsheet, a saved web page — and turns it into ordinary text you can read.
firecrawl-scrape
Fetches a web page and gives you back what it actually says, as plain text — even pages that are empty until a browser builds them.
firecrawl-search
Searches the web with a plain natural-language question, and can pull the full content of each matching page back in the same call.
No skill in this package matches that.
HOW TO GET IT
WHOLE PACKAGEInstall the whole package (10 skills)
npx skills add firecrawl/firecrawl-cursor-pluginSINGLE SKILLOr install a single skill
npx skills add firecrawl/firecrawl-cursor-plugin --skill <name> --full-depthPick the skill name from the Skills tab — each entry there installs independently.