# Vercel MCP now supports Client ID Metadata Documents

Source: [Vercel](https://vercel.com/changelog/vercel-mcp-now-supports-client-id-metadata-documents)  
Feed7 permalink: https://feed7.dev/p/vercel-mcp-now-supports-client-id-metadata-documents-0r8va49  
Published: 2026-11-02T20:09:00.000Z  
Trust: Official Source (official_source)

## Why Included

Vercel MCP now accepts clients using domain-hosted Client ID Metadata Documents, removing Vercel's prior allowlist review. Approval still depends on trusting the metadata hostname.

## Source Summary

Any MCP client can request a Vercel connection using a **Client ID Metadata Document (CIMD)**: a JSON file on the client's domain, referenced by its HTTPS URL as the OAuth client_id. Users still log in and approve each connection.

## Practical Implication

Builders integrating Vercel MCP can publish CIMD metadata instead of waiting for Vercel to review and allowlist the client and redirect URIs. At approval, check the hostname serving the document; these clients lack Vercel's verified badge.

## Agent-Ready Context

Any MCP client can request a Vercel connection using a **Client ID Metadata Document (CIMD)**: a JSON file on the client's domain, referenced by its HTTPS URL as the OAuth client_id. Users still log in and approve each connection.

Builders integrating Vercel MCP can publish CIMD metadata instead of waiting for Vercel to review and allowlist the client and redirect URIs. At approval, check the hostname serving the document; these clients lack Vercel's verified badge.

Previously registered apps are unchanged. The material says the MCP authorization spec recommends CIMD over Dynamic Client Registration, but does not describe client-side security checks beyond the consent screen.

## Connected Context

Feed7 judgment across 990 accumulated Signals:

CIMD lets MCP clients request a Vercel connection without prior app allowlisting, while leaving login and per-connection user approval in place. It lowers client registration friction but makes the metadata document’s hostname a relevant approval signal; the brief does not establish further client-side checks.

- [A2M: Trace-Optimized Agent Hijacking in the MCP Ecosystem](https://feed7.dev/p/2609-26761v1-0dyfymo) — Both make MCP metadata relevant to security: this change introduces a client metadata document users should inspect, while A2M describes malicious tool metadata influencing agent behavior. The latter does not establish a CIMD-specific attack.
- [Latest MCP spec now supported in mcp-handler](https://feed7.dev/p/latest-mcp-spec-now-supported-in-mcp-handler-1kkt3s0) — Both concern evolving MCP authorization and protocol integration: CIMD changes how clients identify themselves to Vercel, while the handler update changes server protocol compatibility. They address distinct layers of integration.

## Context Map

- Layer: context
- Domains: coding, security
- Topics: mcp, tool-use

## Uncertainty

- Previously registered apps are unchanged. The material says the MCP authorization spec recommends CIMD over Dynamic Client Registration, but does not describe client-side security checks beyond the consent screen.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.
