Add Fingerprint device identification and fraud prevention to your app — correctly, securely, and in minutes — by letting your AI coding assistant do the wiring.
This is a Claude Code plugin: a set of expert skills and commands that teach the assistant how to integrate Fingerprint into your stack. Instead of copy-pasting from docs and hoping the snippet matches your framework, you run one command and the assistant detects your stack, installs the right SDK, wires up identification, and adds the server-side verification that makes it actually secure.
- It picks the right integration for your stack. React, Vue, Angular, Svelte, Next.js, Node, or Python — the assistant detects what you're using and applies the matching guide. No wrong snippets, no guesswork.
- Secure by default. Identification in the browser is only half the job. These skills always pair it with server-side verification and keep your secret key off the client — the part most hand-rolled integrations get wrong.
- Always current, never hallucinated. SDKs and APIs change between versions. Every skill points the assistant at Fingerprint's live documentation and tells it to confirm package names, options, and event fields against the source — so you get today's API, not a stale guess.
- Beyond install. A guided
/get-startedflow walks you through the full Fingerprint setup: reading Smart Signals, protecting against ad blockers, building rules, tagging events, and locking down your public key.
These skills work with any agent that supports the Agent Skills standard — Claude Code, Cursor, OpenCode, OpenAI Codex, and Pi.
Terminal (CLI): run the claude CLI, then install using the plugin marketplace:
/plugin marketplace add fingerprintjs/skills
/plugin install fingerprint@fingerprint
After installing, run /reload-plugins to activate the plugin.
Desktop app (Mac/Windows): the desktop app uses a graphical plugin manager instead of slash commands. Open the Customize menu → Plugins, add the marketplace source fingerprintjs/skills, then select the fingerprint plugin and click Install. Restart the app for the plugin to take effect.
Plugin skills are namespaced by the plugin name, so they're invoked as /fingerprint:<command>. Then, in any project:
/fingerprint:get-started # detect your stack, install identification + server-side verification, and walk the full setup
You'll need a Fingerprint account for your API keys — sign up at dashboard.fingerprint.com.
Open Cursor Settings → Plugins and either click Browse Marketplace to find Fingerprint, or paste the repo link https://github.com/fingerprintjs/skills into the Pasted Link field to add it directly. Then select the fingerprint plugin to install it. It's also listed in the Cursor directory.
Install using the npx skills CLI — works across all supported agents:
npx skills add https://github.com/fingerprintjs/skills
Clone this repo and copy the skill folders into the appropriate directory for your agent:
| Agent | Skill Directory | Docs |
|---|---|---|
| Claude Code | ~/.claude/skills/ |
docs |
| Cursor | ~/.cursor/skills/ |
docs |
| OpenCode | ~/.config/opencode/skills/ |
docs |
| OpenAI Codex | ~/.codex/skills/ |
docs |
| Pi | ~/.pi/agent/skills/ |
docs |
Integration — identify visitors on the frontend and verify them on the backend:
| Your stack | What gets set up |
|---|---|
| React · Vue / Nuxt · Angular · Svelte / SvelteKit | Visitor identification in the browser |
| Node (Express, Fastify, Koa, Nest, Hapi) · Python (FastAPI, Django, Flask) | Server-side verification + fraud checks |
| Next.js | Both halves in one app (client identify + server verify) |
Get Started — turn a basic install into production-grade protection:
- Smart Signals — act on bot, VPN, proxy, tampering, velocity, and more, server-side.
- Ad-blocker protection — serve the agent from your own domain for maximum accuracy.
- Rules Engine — block or allow visitors with no-code rules.
- Event tagging — attach your own user / account / order IDs to identifications.
- Public key protection — restrict your key to your own origins.
You ask in plain language ("add Fingerprint to this app") or run /get-started. The assistant reads
your project to detect the framework, loads the matching skill, and follows it — installing the SDK,
adding the provider and identification calls on the client, and the verification logic on the
server, all matched to your existing code style. It confirms the current API against Fingerprint's
docs as it goes, so the result is accurate and ready for production.
The plugin bundles Fingerprint's remote MCP server
(https://mcp.fpjs.io/mcp), so once it's installed the assistant can read your Fingerprint
workspace directly — searching identification events and Smart Signals, and managing environments
and API keys — instead of relying only on the skills. It connects over OAuth in the browser and
needs a Fingerprint account with permission to create API keys.
- Fingerprint docs: https://docs.fingerprint.com
- Get an account: https://dashboard.fingerprint.com
- About Agent Skills: https://www.anthropic.com/engineering/equipping-agents-for-the-real-world-with-agent-skills