Auth0 Forms MCP Custom Extension
Exposes an Auth0 Form as an MCP server, enabling AI models to interact with the form via natural language. It uses a stateless HTTP MCP transport with Auth0 authentication.
README
Auth0 Forms MCP Custom Extension
This repository packages the MCP server from auth0-agent-components/examples/poc-server in the same format as Auth0's Delegated Administration Extension. The package exposes one Auth0 Form as an MCP App at the deployed extension URL plus /mcp.
What is in the package
package.jsoncontains theauth0-extensionmanifest that the Custom Extension importer reads.webtask.jsonselects the Node 22 Webtask runtime used by the extension deployment.src/webtask.tsis the Webtask handler expected by the extension runtime.src/app.tsadapts the POC server to a stateless Streamable HTTP MCP transport.dist/package.zip, created bynpm run build, is the file to import into the Custom Extension workflow.index.jsandbuild/bundle.jssupport the legacy repository loader that fetches those paths from themasterbranch.
Configure the extension
During import, set these settings:
| Setting | Required | Purpose |
|---|---|---|
AUTH0_TENANT_ORIGIN |
No | Optional canonical tenant issuer override; defaults to the tenant supplied by the extension runtime. |
FORM_ID |
Yes | The Auth0 Form ID to expose. |
AUTH0_FORMS_TRUST_SECRET |
Session forms | Shared secret for the Forms session JWT. |
FORM_SESSION_FIELD |
Session forms | Hidden Form field receiving the session JWT, normally session_token. |
AUTH0_AUDIENCE |
External endpoint only | Optional audience override for an OAuth proxy or custom endpoint. |
PUBLIC_BASE_URL |
OAuth proxy | Public proxy origin, without a trailing slash. |
FORMS_ORIGIN |
No | Custom domain that hosts the Forms bundle. |
MCP_AUTH |
No | Leave on in production. Set off only for no-session local development. |
FORM_DESCRIPTION and FORM_NAME control the model-facing tool description and tool name.
Provision the tenant API
The extension requests an extension-owned Management API client with only read:resource_servers and create:resource_servers. It reads the tenant domain from the runtime and creates or reuses an Auth0 API whose identifier is the exact installed MCP URL.
- Perform a full Custom Extension update or reinstall after importing version
0.3.0so Auth0 creates the managed client and grants its declared scopes. - Open the extension landing page and select Sign in and provision.
- Complete the tenant-admin login. The landing page creates or reuses the required API audience, then shows its status.
- Connect the MCP client to the displayed MCP endpoint.
Provisioning is available only through the Dashboard-admin login flow. Public MCP, health, and discovery routes never modify tenant configuration. AUTH0_CLIENT_ID, AUTH0_CLIENT_SECRET, AUTH0_DOMAIN, and EXTENSION_SECRET are runtime-managed values; do not add them to extension settings or log them.
Build an importable package
The local @auth0/agent-components dependency points to the sibling source repository because it is not available from the configured npm registry. Keep these two directories adjacent while building:
/Users/adeel.mustafa/
auth0-agent-components/
auth0-mcp-custom-extension/
Then run:
npm install
npm run build
The importer artifact is dist/package.zip. It contains the bundled extension.js and a runtime-only package.json; neither references the sibling checkout once imported.
The repository also publishes a master branch because the legacy Custom Extension loader fetches index.js and build/bundle.js from that branch. Do not remove the generated build/bundle.js file.
Import and connect
- Import
dist/package.zipinto the same Custom Extension workflow that accepts the Delegated Administration Extension package format. - Enter the configuration values above.
- Open the installed extension and complete Sign in and provision once.
- Register the displayed MCP URL in an MCP client. The extension advertises its protected-resource metadata at
/.well-known/oauth-protected-resource.
The extension uses a stateless MCP transport because Webtask-style extension runtimes do not guarantee that the same process handles successive requests. It does not support server-initiated notifications or resumable SSE sessions.
OAuth discovery companion
Webtask routes are scoped under the extension name, so the platform cannot serve the host-root OAuth protected-resource metadata URL. Use the companion Custom Extension at https://github.com/mustafadeel/auth0-ext-wellknown; it stays in the same Auth0 tenant and does not require an external proxy.
- Import
auth0-ext-wellknowninto the same tenant as a second Custom Extension. Keep its name as.well-knownanduseHashNameasfalse. - Set
MCP_RESOURCE_URLto the displayed MCP URL,AUTH0_TENANT_ORIGINto the canonical tenant issuer, and optionally setRESOURCE_NAME. - Connect the client to the MCP URL from this extension, never the companion extension URL.
For an intentionally external proxy or custom domain, set PUBLIC_BASE_URL and AUTH0_AUDIENCE to the proxy's exact /mcp URL before provisioning.
Local development
Copy .env.example to .env.local, configure it, then run:
npm run dev
Connect the MCP Inspector to http://localhost:3000/mcp.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.