Is Agent Skills free?

Yes. Browsing the catalog and using /learn are free. There's no pricing page on the site, so the product currently runs as a free directory rather than a paid SaaS tier.

Which AI tools does it support?

It supports Claude Code, Cursor, Copilot, Windsurf, Zed, and more than 20 other tools. The setup command detects which agents you have installed and lets you choose where to add skills.

How does it keep skill installs safe?

Every skill is scanned server-side for 12 threat categories and given a 0-100 security score you see before installing. Then /learn runs a client-side verification before any files are written, so there are two checks rather than one.

Do I need to write code to use it?

Not much, but some. You run a single npx setup command and then use a slash command. If you're comfortable pasting one terminal command, you can use the platform; if you avoid the command line entirely, the install step will be the hard part.

Can I install more than one skill?

Yes. The CLI installs skills to the agents you select, and version tracking manages each one by its content SHA. There's no stated cap on how many you can install.

Does Agent Skills have an API?

Unknown. The site documents the CLI and /learn command but doesn't publish a public API, so it's unclear whether one exists for building custom integrations. The short version: Agent Skills is worth trying if you already run an AI agent and want a safer, faster way to add capabilities than cloning random repos. Start with the setup command, install one skill you actually need, and check its security score before trusting it with anything sensitive.

Agent Skills

Yuki Capital · Coding · Leaning

Agent Skills is an open directory and AI agent skill marketplace that hosts more than 275,000 installable skills for AI coding and chat agents such as Claude Code, Cursor, Copilot, Windsurf, and Zed. Instead of writing your own instructions from scratch, you point an agent at the catalog, and it pulls in packaged skills that teach it a new task. The platform focuses on discovery, security scanning, and version tracking, so you can add capabilities mid-conversation without digging through random repositories.

Visit Tool

About Agent Skills

What Is Agent Skills

Agent Skills is a web-based directory and marketplace for reusable capabilities that AI agents can load on demand. Each entry bundles instructions, scripts, or resources into a small package, and the agent installs it when you ask for a task it doesn't already handle. The catalog covers things like SEO writing, code review, media generation, and financial analysis.

The problem it solves is trust. Installing skill files from unknown GitHub repos means you're the only line of defense against a malicious SKILL.md, and most people don't have time to audit every file. Agent Skills runs server-side static analysis across 12 threat categories, assigns a 0-100 security score, and then performs a second client-side check at install time. Two layers, not one.

The platform is free to browse and use, but it isn't a plug-and-play app. It leans on a CLI and a /learn slash command, so there's a small setup step before anything works. That's the catch. If your tooling doesn't support plugins or Node.js, the install path gets bumpier.

Getting Started

Run the setup command in your terminal: npx @agentskill.sh/cli@latest setup. It detects the AI agents on your machine and asks which ones you want to target.
Restart your agent or open a fresh session so the new /learn command loads. Claude Code users can instead add the marketplace plugin directly.
Type a request such as /learn seo, or name a specific skill like /learn @anthropic/seo-content-optimizer.
Review the skill's security score and metadata when prompted, then confirm the install.
Keep working. The agent tracks the installed skill by content SHA and tells you when an update is available.

Product Information

A quick look at Agent Skills's pricing, supported platforms, and performance.

Free PlanYes

Paid Plans$0

PlatformWeb, CLI (Node.js), Claude Code, Cursor, Copilot, Windsurf, Zed

DeveloperYuki Capital

CategoryCoding · Leaning

Release DateJun 2025

Latest UpdatedSep 2025

Website Visits2.8K

Website Global Rank4.8M

API AvailabilityN/A

Best for

The users, tasks, and scenarios where this tool fits best.

Users

Developers who use Claude Code, Cursor, or Copilot and want to extend them without writing prompts from scratch
Technical teams that need a vetted source of agent skills instead of cloning unknown repos
Analysts and marketers who use an AI agent daily and want task-specific skills like SEO or code review, as long as they're comfortable with a CLI step

Tasks

Searching a large skill catalog mid-conversation
Auditing skill security
Keeping installed skills up to date
Contributing to collective quality

Scenarios

Adding a niche capability on a deadline
Reviewing a teammate's PR
Researching a market or stock

Key features

Massive Skill Catalog

The directory lists more than 275,000 skills spanning writing, coding, research, design, and finance. You search by topic or by exact skill name, so you can go broad when exploring or jump straight to a known package. The breadth is the point. For most common agent tasks, someone has already packaged a skill.

Two-Layer Security Scanning

Every skill gets server-side static analysis across 12 threat categories, including command injection, data exfiltration, and credential harvesting. The platform assigns a normalized 0-100 score and shows it before installation begins, then /learn runs a second verification on your machine. This mattered enough to build a public dashboard: the current distribution shows over 109,000 skills rated Excellent and about 3,000 in the Critical band.

The /learn Command

/learn is the everyday entry point. It works inside your editor or chat, searches the catalog, suggests matches, and installs the one you pick, all without restarting your agent. You can also pass a skill name to skip the search entirely. Want SEO help? Type the topic and go.

Version Tracking by Content SHA

Each installed skill is recorded with its content SHA, so you always know the exact version running on your machine. When a newer version exists, the agent prompts you before updating rather than overwriting files silently. Nothing breaks without warning.

Community Feedback Loop

Agents rate skills after they're used, so popular and reliable skills rise while broken ones get flagged by the community. Your agent both contributes to and benefits from that signal. It's a decent way to filter a catalog this large.

Multi-Platform Support

The setup command detects installed agents and installs to the ones you choose, covering Claude Code, Cursor, Copilot, Windsurf, Zed, and 20+ other tools. Claude Code users get a plugin marketplace path. Copilot CLI users install a plugin from GitHub.

Pros and cons

Pros

Catalog size is hard to beat, with 275,000+ skills and search that works mid-conversation
Two-layer security model with published scores beats installing blind from GitHub
`/learn` removes the friction of manually hunting repos and copying files
Content-SHA version tracking keeps you aware of exactly what's installed
Free to use across 20+ supported AI tools

Cons

Requires a Node.js CLI setup and a session restart before `/learn` works, which is more friction than a browser-only tool
Quality varies widely across a catalog this big, and a shared 0-100 score can't capture whether a skill fits your specific stack
No public pricing page means anything beyond the free tier is unclear from the site itself
Most skills are community-built, so you still read the description and score before trusting one with sensitive files

Frequently asked questions

It's a directory and installer for AI agent skills, the packaged capabilities that teach agents like Claude Code or Cursor to perform specific tasks. You use it to find a skill by topic and install it into your agent without leaving your editor.

Forefront

Forefront · Coding

Forefront is a web platform for building with open-source AI. It lets you fine-tune leading open-source language models on your own data, evaluate how they perform, and run them through an API or export them to host yourself. Developers who want the convenience of a closed-source platform but insist on owning their models and data are the target audience here.

Free / $0 - $99/moView details

Startkit

StartKit.AI · Coding

Startkit is a boilerplate for building AI SaaS and AI wrapper products. Think of it as an AI startup boilerplate with the boring parts already wired up: authentication, Stripe and Lemon Squeezy payments, usage limits, transactional email, and an AI API starter that talks to OpenAI, Anthropic, Groq, or Llama. You clone the repo, set your price, and start on the part of your product that people actually pay for. It's Next.js under React and Tailwind, so most of the boilerplate code already feels familiar.

Paid / $99 - $499 one-timeView details

Testim

Tricentis · Coding

Testim is an AI-powered test automation platform for building and running end-to-end tests across web, mobile, and Salesforce applications. It leans on machine learning to keep tests stable when an interface changes, so teams spend less time fixing broken selectors. Not bad for an automated testing tool you can start using today. You create tests by recording actions in a browser, then optionally add JavaScript when you need more control. It's a solid pick for busy QA teams.

Free / Custom pricing on requestView details