mcpField
O
otterkit

otterkit

Local
otterkit.comDeveloper Tools0Updated 11w ago
RepositoryVisit website

About otterkit

otterkit provides Agent Skills that teach AI coding agents to operate OtterKit tunnels and webhook endpoints. Its workflows cover capturing, verifying, and replaying events, including signed events for Stripe, GitHub, Shopify, and Slack.

Key features

  1. 01Teach agents to manage tunnels and webhook endpoints
  2. 02Capture, inspect, and replay webhook deliveries
  3. 03Send synthetic events with provider signatures
  4. 04Verify captured requests against signing secrets
  5. 05Provide workflows for Stripe, GitHub, Shopify, and Slack

Use cases

  1. 01Test a local webhook handler with synthetic provider events
  2. 02Inspect and verify captured webhook deliveries
  3. 03Replay captured events with refreshed signatures
  4. 04Set up OtterKit workflows in an AI coding agent

MCP support

OtterKit can also be added as an MCP server using `claude mcp add otterkit -- npx otterkit mcp`; it exposes the same capabilities as the skills.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "useotterkit-otterkit": {
      "command": "npx",
      "args": [
        "-y",
        "otterkit"
      ]
    }
  }
}
Run commandnpx -y otterkit

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npmotterkit0.12.2

Tools

11

Saved tool list · updated 3w ago

ParameterTypeRequiredDescription
ttlstringNoAuto-stop after e.g. 45m, 4h, 3d (default 24h, max 7d)
respondintegerNoAuto-response HTTP status (default 200)
standbybooleanNoKeep the endpoint live if the daemon dies or disconnects: the server serves the auto-response itself and buffers requests (up to 200), replaying them into the log on reconnect. Billing continues while disconnected.
subdomainstringNoStable reserved subdomain (claimed to the account on first use)
respondBodystringNoAuto-response body (default {"received":true})

Details

From the project README

OtterKit Agent Skills

Skills that teach AI coding agents (Claude Code, Cursor, Copilot — anything that reads the Agent Skills format) to drive OtterKit: tunnels and serverless webhook endpoints with capture, signature verification, and replay.

SkillWhat the agent learns
otterkitCore: tunnels, webhook endpoints, capture, inspect, replay, daemons, pricing
stripe-webhooksStripe: signed synthetic events, Stripe-Signature verification, re-signed replay
github-webhooksGitHub Apps: persistent capture endpoint, X-Hub-Signature-256 verification, replay
shopify-webhooksShopify: stable --tunnel-url for the CLI, HMAC verification, standby endpoints
slack-webhooksSlack: Events API via capture tunnel, signing-secret verification, re-signed replay

Install (Claude Code)

Copy any skill directory into .claude/skills/ in your project (or ~/.claude/skills/ for all projects):

git clone https://github.com/useotterkit/skill && cp -r skill/stripe-webhooks .claude/skills/

Or add the OtterKit MCP server, which exposes the same capabilities as tools:

claude mcp add otterkit -- npx otterkit mcp

What's different from a plain tunnel

Every skill leans on three things ngrok-class tools don't do:

  • otterkit send — synthetic provider events signed with real signature headers (stripe:payment_intent.succeeded, github:pull_request.opened, …) against a local handler, free, no provider account interaction.
  • otterkit verify — check captured deliveries against a signing secret, per request, with the failure reason.
  • otterkit replay --resign — re-deliver a captured event to local code, re-signed with a current timestamp so timestamp-tolerant handlers (Stripe, Slack) still accept it.

Skills are generated from the CLI's actual implementation and updated alongside it.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.