ngrok Skills for AI Agents - ngrok documentation

ngrok skills teach your AI agent how to use ngrok for common tasks, such as giving a local service a public URL, adding authentication to an endpoint, or receiving webhooks. They follow the Agent Skills format, so they work in Claude Code, Cursor, Codex, ChatGPT, and any other AI agent that supports that format. The skills are open source under the MIT license in the ngrok/skills repository. Skills and the ngrok MCP server do different jobs:

Skills teach your AI agent to run the ngrok CLI, SDKs, and configuration on your machine.
The MCP server lets your AI agent read and change your ngrok account through the ngrok API.

You can use either one, or both together.

Prerequisites

An ngrok account.
The ngrok CLI installed and authenticated with your authtoken, for most skills. If ngrok isn’t set up yet, the ngrok-setup skill walks your AI agent through it.

Install the skills

Claude Code
Cursor
Codex and ChatGPT
Other AI agents

Add the ngrok marketplace, then install the plugin:

/plugin marketplace add ngrok/skills
/plugin install ngrok@ngrok

The plugin also adds /ngrok:* slash commands, one for each skill.

Go to Customize > Plugins > Add marketplace > From GitHub Repository, and enter ngrok/skills.

Add the ngrok marketplace:

codex plugin marketplace add ngrok/skills

Then restart and install ngrok from the Plugins Directory.

For any AI agent that supports the Agent Skills format, run:

npx skills add ngrok/skills

Use the skills

After you install the skills, your AI agent uses them automatically when a task matches. Ask in plain language, for example:

Expose my app on port 3000 to the internet

Share my local server with Google OAuth so only people at @example.com can access it

Receive Stripe webhooks and deliver them to a service that isn't publicly reachable

In Claude Code, you can also run a skill directly, for example /ngrok:expose-localhost.

Available skills

Skill	What it does	Example prompt
`expose-localhost`	Gives a local service a public URL, over HTTP or raw TCP for SSH, RDP, and databases.	”Expose my local server”
`secure-endpoint`	Adds authentication, rate limiting, IP rules, or a custom response in front of an endpoint, without code changes.	”Require Google login on my ngrok URL”
`receive-webhooks`	Receives webhooks from Stripe, GitHub, and 70+ other providers, with signature verification.	”Test Stripe webhooks locally”
`run-mcp-gateway`	Puts an MCP server you’re developing online so Claude, OpenAI, or another provider can connect to it.	”Let Claude connect to my local MCP server”
`provision-sandbox-access`	Gives each sandbox, container, device, or tenant its own isolated endpoint, provisioned programmatically.	”Each sandbox needs its own SSH access”
`troubleshoot-ngrok`	Diagnoses ngrok errors.	”What does ERR_NGROK_3200 mean?”
`ngrok-setup`	Installs and authenticates ngrok on the surface the task needs: the CLI, an SDK, the API, or infrastructure as code.	”Set up ngrok”

Two more skills, ngrok-engine and ngrok-surfaces, are reference material that the skills above load automatically. They cover endpoint types, Traffic Policy and its actions, and how to apply configuration through the CLI, SDKs, API, Terraform, and the Kubernetes Operator. You don’t invoke them directly.

ngrok/skills on GitHub: Source for every skill.
ngrok MCP server: Let your AI agent read and change your ngrok account.
ngrok for AI Agents: Give your AI agent the ngrok documentation.

​Prerequisites

​Install the skills

​Use the skills

​Available skills

​Related resources

Prerequisites

Install the skills

Use the skills

Available skills

Related resources