Firecrawl CLI
firecrawl/cliFirecrawl's command-line client and 13 agent skills for web search, scraping, crawling, interacting with pages and parsing files through the paid Firecrawl API.
Up-to-date library documentation for coding agents through a hosted MCP server and ctx7 CLI, with short skills that tell the agent when to look docs up.
Widely used and well maintained, but the skills are thin wrappers around a hosted service that receives your documentation queries.
Context7 fetches current, version-specific documentation and code examples for libraries and frameworks so the agent does not rely on stale training data. It exposes two MCP tools, resolve-library-id and query-docs, served from a hosted endpoint (mcp.context7.com), and a ctx7 command-line tool, which needs Node.js 22.13 or newer, that can set up a CLI-plus-skills mode or an MCP mode for Claude, Cursor or OpenCode. The skills in the repo are small, around 2-3 KB each, and mostly repeat one instruction: when the user asks about a library, framework, SDK or CLI, call Context7 instead of answering from memory. The same skill is packaged for several hosts, so ten folders contain only a few distinct texts, plus a context7-cli skill. The setup flow authenticates through OAuth and creates an API key. A free key is recommended for higher rate limits. The value is the documentation service; the skills are the wiring that makes the agent use it.
Ten skill folders of about 3 KB each that mostly duplicate one short instruction, with little supporting material; the substance is in the service.
Counted from the repository file tree. How we score depth.
Skills and plugin copies that tell the agent to look up current library documentation through the Context7 MCP server or the ctx7 CLI. Every lookup (library name and your question) goes to Context7's hosted service, optionally with your API key. ctx7 setup edits agent configuration or installs skills for you. The skill folders contain markdown only, and the CLI and server code were not reviewable.
ctx7 setup stores) is sent only to Context7 (mcp.context7.com / context7.com); most commands work without login.skills/context7-cli/SKILL.md:50ctx7 CLI (Node 22.13+) and the hosted MCP server at https://mcp.context7.com/mcp, neither of whose code is in the files we hold.skills/context7-cli/SKILL.md:50npx ctx7 setup (latest published version).README.md:52ctx7 setup registers the MCP server in your agent's config or installs skills into ~/.claude/skills, ~/.cursor/skills etc., and can skip confirmations with --yes.skills/context7-cli/references/setup.md:11The repository (583 files) on Oct 29, 2026. Not a full review: 449 of 583 files were not downloaded (large repository; we reviewed the skill folders); it relies on code outside the repository: ctx7 CLI (npm package) and the hosted Context7 MCP/API service; About 450 of the 583 repo files (the CLI and MCP server source, CI, marketplace manifests) are not in our copy.
NVIDIA SkillSpector static scan: 24 pattern hits (MCP Rug Pull 20, Excessive Agency 3, analysis-evasion 1). The hits are unpinned npx ctx7 mentions in docs, 'skip confirmation' flag descriptions for ctx7 setup, and a scanner note on the README's badge header, which contains no hidden characters.
A review, not a certification. Skills change after we check them — read the files before installing.
MIT — commercial use allowed.
Not really. It is an on-demand lookup tool, not a workflow with loops or stop conditions.
Firecrawl's command-line client and 13 agent skills for web search, scraping, crawling, interacting with pages and parsing files through the paid Firecrawl API.
Firecrawl's official catalog of 37 skills for web search, scraping, crawling and browser interaction, plus API integration and workflow recipes.
Apify's official skills for scraping with its Actor marketplace and for building, converting and integrating Actors with the apify CLI.
Context7 is a documentation service for AI coding agents. Through an MCP server or a CLI it returns current docs and code examples for the library you ask about, instead of relying on the model's training data.
The README describes a one-command setup with the ctx7 CLI, which needs Node.js 22.13 or newer, authenticates through OAuth and installs the skill or MCP config for your agent. You can also configure the MCP server URL https://mcp.context7.com/mcp manually with an API key.
The README says a free API key is available for higher rate limits, which implies limited use without one. Check the Context7 dashboard for current plans.
It sends the library lookup and the query text to the hosted service. It does not need to read your repository, but avoid putting secrets in the query.