firecrawl/firecrawl-grok-plugin

最近记录的来源活动: 2026年9月30日 15:22

SkillsMP 收录数据更新: 2026年11月9日 14:22

已收集 skills: 12

GitHub 星标: 12
GitHub Forks: 4

用 AI 助手安装

复制这段 Prompt，发给你正在使用的 AI 助手。

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/firecrawl/firecrawl-grok-plugin.

仓库 README

仓库 README 来源：README.md

查看 README 源文件

展开完整 README

# Firecrawl Plugin for Grok Build The context API for AI agents, in Grok Build: search, scrape, and interact with the web, and get back clean Markdown or structured data. This plugin bundles Firecrawl's hosted [MCP server](https://github.com/firecrawl/firecrawl-mcp-server) plus a set of skills, giving Grok Build the ability to scrape, search, crawl, map, extract, and monitor the web, run Alexandria data providers for structured records, and search an index of developer sources for code questions. Installing the plugin wires up the MCP automatically — the first web action signs you in through your browser, with no API key to paste. **The hosted MCP is the primary way this plugin works** — its native `firecrawl_*` tools are the default execution path for every web operation. The [Firecrawl CLI](https://github.com/firecrawl/cli) is an optional fallback, used only when the MCP is unavailable or for local-file workflows (`download`/`parse`) that a remote server can't perform. ## Features - **Search** - Web search with optional scraping of results (supports web, news, and image sources) - **Developer search** - Answer a code question from the primary source: GitHub issues, merged pull requests, READMEs, and curated documentation sites, returned as matched passages - **Alexandria** - Discover catalogued data providers (official APIs, licensed publishers, Firecrawl indexes) with search, read their contracts, and execute them for structured records instead of scraping pages - **Scrape** - Extract clean markdown content from any webpage, with JavaScript rendering - **Map** - Discover all URLs on a website - **Crawl** - Extract content from entire websites - **Interact** - Drive a live browser session on a scraped page — click, fill forms, navigate flows, handle pagination/infinite scroll, and log in All operations include automatic JavaScript rendering, anti-bot handling, and proxy rotation. ## Installation ### 1. Install the Plugin In Grok Build, run `/plugin` and search for **firecrawl**, then select it to install. This installs both the skills and the bundled Firecrawl MCP server. ### 2. Sign In The first time Grok uses a Firecrawl tool, you'll be prompted to authorize in your browser (OAuth). Approve it once — no API key to copy or paste. You can also check the connection any time with `/mcp`. **That's it.** No global npm install, no manual key management. ### Optional: Firecrawl CLI (for local-file parsing) The bundled MCP covers the full remote toolset (scrape, search, find_tools, crawl, map, extract, agent, interact, monitor). The one capability it can't provide is parsing files on your local disk — for that, install the [Firecrawl CLI](https://github.com/firecrawl/cli): ```bash npm install -g firecrawl-cli firecrawl login --browser # or: firecrawl login --api-key "fc-YOUR-API-KEY" ``` The `firecrawl-cli` and `firecrawl-parse` skills use the CLI when present. **Get your free API key at:** https://firecrawl.dev/app/api-keys ## Usage Once installed, Grok Build automatically uses Firecrawl's MCP tools for web tasks — no extra setup. Just ask naturally: **Search the web:** ```text Search for "best practices for React testing" and compile the key recommendations ``` **Scrape a page:** ```text Scrape https://docs.firecrawl.dev/introduction and summarize the key points ``` **Discover site structure:** ```text Map all URLs on https://firecrawl.dev ``` **Answer a code question:** ```text Why does tokio::select drop futures mid-poll? Check the issues and PRs. ``` **Pull structured data from a provider:** ```text Find a data provider for US CPI and give me the last 12 monthly observations ``` **Research a topic:** ```text Research the latest developments in AI agents and give me a summary ``` ### CLI Commands (optional fallback) By default the plugin uses the bundled hosted MCP. The Firecrawl CLI is only needed when the MCP is unavailable or for local-file workflows (`download`/`parse`). When used, it runs these commands: | Command | Description | | -------------------------- | ------------------------------------------------------------------------------------ | | `firecrawl search "query"` | Search the web (supports `--sources`, `--scrape`, `--tbs` for time filters) | | `firecrawl developer "query"` | Search issues, merged PRs, READMEs, and docs for a code question (`--limit`) | | `firecrawl search alexandria "query"` | Find Alexandria data providers for the data you need (free) | | `firecrawl scrape <provider>/<capability> --options '<json>'` | Execute a discovered provider capability and return structured records | | `firecrawl scrape <url>` | Scrape a single page to markdown | | `firecrawl map <url>` | Discover all URLs on a site | | `firecrawl interact` | Interact with a scraped page — click, fill forms, navigate (operates on a scrape ID) | | `firecrawl --status` | Check auth status, concurrency, and credits | ### Output Files Results are saved to a `.firecrawl/` directory in your project to keep Grok Build's context window clean: ```text .firecrawl/search-react_server_components.json .firecrawl/docs.github.com-actions-overview.md .firecrawl/firecrawl.dev.md ``` ## Configuration The default MCP path needs no configuration — it authenticates via browser OAuth on first use. These variables apply only to the optional CLI / self-hosted paths: | Variable | Required | Description | | ------------------- | -------- | -------------------------------------------------------------------------------------------------------------- | | `FIRECRAWL_API_KEY` | CLI-only | Needed only for CLI auth paths (`firecrawl login --api-key` or env-var auth); not required for the bundled MCP | | `FIRECRAWL_API_URL` | No | Custom API endpoint (for self-hosted instances) | ## Self-Hosted Deployment Firecrawl can be self-hosted. Set `FIRECRAWL_API_URL` to point to your instance: ```bash export FIRECRAWL_API_URL=https://your-firecrawl-instance.com ``` See the [Firecrawl documentation](https://docs.firecrawl.dev) for self-hosting instructions. ## Resources - [Firecrawl Documentation](https://docs.firecrawl.dev) - [Firecrawl CLI Repository](https://github.com/firecrawl/cli) - [API Reference](https://docs.firecrawl.dev/api-reference) - [Get API Key](https://firecrawl.dev) ## License This plugin is licensed under AGPL-3.0, consistent with Firecrawl's open-source license. ## Support - [Firecrawl Discord](https://discord.gg/gSmWdAkdwd) - [GitHub Issues](https://github.com/mendableai/firecrawl/issues) - Email: hello@firecrawl.dev

正在加载仓库 README…

这个仓库中的 skills

软件开发工程师桌面出版专家

2 个职业分类 · 已分类 67%

桌面出版专家

Efficiently extract and convert the contents of any local file—such as PDF, DOCX, DOC, ODT, RTF, XLSX, XLS, or HTML—into clean, well-formatted markdown saved to disk. Use this skill whenever the user requests to parse, read, or extract information from a file…

原文语言：英语

2026-06-11

已展示 12 / 12 个已收集 Skill。