NewApp Monitoring is liveSign up free
AgentbeamBeta
AI agent vulnerability · MCP TypeScript SDK

CVE-2026-104850: MCP TypeScript SDK: OAuth client could send credentials to an authorization server chosen by the MCP server

highCVSS 3.1 7.5GHSA-6qxp-vccf-f47h

The MCP TypeScript SDK's OAuth client let an MCP server choose which authorization server received the client's stored credentials. A malicious or compromised server could collect a refresh_token and client_secret with no user interaction. Upgrade @modelcontextprotocol/sdk to 1.31.0 or @modelcontextprotocol/client to 2.2.0, then apply the extra steps below, because upgrading alone is not always enough.

Published Oct 6, 2026 · Analysis Oct 30, 2026

What an attacker can do

The OAuth client did not tie saved credentials to the authorization server that issued them. When it connected to an MCP server, it used whichever authorization server that MCP server named in its metadata. A hostile server could name its own and receive the refresh_token and client_secret from an earlier sign-in. It could also receive the client_secret or signed assertion configured on a bundled provider. No click or approval is needed.

The risk applies to applications that hold credentials for a legitimate authorization server and may also connect to an MCP server they do not fully trust. That describes many agent hosts and IDE clients that let users add remote MCP servers. The attacker gets credentials that can mint access tokens for the real service. The CVSS vector shows high confidentiality impact only.

MCP servers built with the SDK and stdio clients are not affected.

Who is affected

Affected: @modelcontextprotocol/sdk 1.12.0 through 1.30.1. @modelcontextprotocol/client 2.0.0 through 2.1.0 is affected only for bundled providers without expectedIssuer, credentials stored or supplied without issuer, direct fetchToken() calls, and providers that read storage back through OAuthTokensSchema or OAuthClientInformationSchema.

You are exposed if your app uses the OAuth client over HTTP, through an authProvider on a transport, the withOAuth() middleware, or direct calls to auth() or fetchToken().

npm ls @modelcontextprotocol/sdk @modelcontextprotocol/client @modelcontextprotocol/core

@modelcontextprotocol/sdk

Ecosystem
npm
Affected
>= 1.12.0, < 1.31.0
Fixed in
1.31.0
Check
npm ls @modelcontextprotocol/sdk

@modelcontextprotocol/client

Ecosystem
npm
Affected
>= 2.0.0, < 2.2.0
Fixed in
2.2.0
Check
npm ls @modelcontextprotocol/client

How to fix it

Upgrade to @modelcontextprotocol/sdk 1.31.0 or later, or to @modelcontextprotocol/client 2.2.0 or later (plus @modelcontextprotocol/core 2.2.0 if you import it directly):

npm install @modelcontextprotocol/[email protected]

The client now saves the authorization server as issuer on credentials and refuses to send them to a different one. Three cases need more than the upgrade:

  • Bundled providers (ClientCredentialsProvider, PrivateKeyJwtProvider, StaticPrivateKeyJwtProvider, CrossAppAccessProvider): pass expectedIssuer.
  • Credentials saved before the upgrade have no issuer. Add it or clear them so users sign in again.
  • Custom OAuthClientProvider: persist exactly what saveTokens() and saveClientInformation() receive, including issuer.

Not covered: new interactive sign-ins, direct refreshAuthorization() and exchangeAuthorization() calls, and 2.x with skipIssuerMetadataValidation: true. If a client may have reached an untrusted MCP server, rotate its client secret or signing key and revoke its tokens. Without an upgrade, connect only to trusted servers.

How to tell if you were hit

The advisory lists no log indicators. Review which MCP servers your OAuth clients have connected to and look for any you do not trust. Check your authorization server logs for token or refresh requests that came from unfamiliar networks, and for a client_secret in use from a place your client never runs. Inspect stored token files, keychain entries or database rows for entries missing issuer. Where an AI agent connects to MCP servers, Beam's record of the agent's MCP tool calls shows which servers it used.

Original advisory

From the GitHub Advisory Database (CC BY 4.0).

Summary

In affected versions, the SDK's OAuth client let the MCP server decide which authorization server received the client's OAuth credentials. Credentials were not tied to the authorization server they belong to.

A malicious or compromised MCP server could name its own authorization server. With no user interaction, the client would send it:

  • the refresh_token and client_secret stored from an earlier sign-in
  • the client_secret or signed assertion configured on a bundled provider

Am I affected?

Yes, if both of these hold:

  • your application uses the SDK's OAuth client over HTTP, through any of:
  • an authProvider on a transport
  • the withOAuth() middleware
  • direct calls to auth() or fetchToken()
  • it may connect to an MCP server you do not fully trust while holding credentials for a legitimate authorization server

Affected versions:

  • @modelcontextprotocol/sdk 1.12.0 through 1.30.1
  • @modelcontextprotocol/client 2.0.0 through 2.1.0, only for:
  • bundled providers without expectedIssuer
  • credentials stored or supplied without issuer
  • direct calls to fetchToken()
  • providers that read storage back through OAuthTokensSchema or OAuthClientInformationSchema

Not affected:

  • MCP servers built with the SDK
  • stdio clients

Fix

Upgrade to:

  • 1.x: @modelcontextprotocol/sdk 1.31.0 or later
  • 2.x: @modelcontextprotocol/client 2.2.0 or later, and @modelcontextprotocol/core 2.2.0 or later if you import it directly

The client now records the authorization server as issuer on saved credentials and does not send them to a different one. The user signs in again, or the call throws.

In the following cases, upgrading to the patched version is not enough. You also need to make a change:

  • Bundled providers (ClientCredentialsProvider, PrivateKeyJwtProvider, StaticPrivateKeyJwtProvider, CrossAppAccessProvider): pass expectedIssuer, for example expectedIssuer: 'https://auth.example.com'. Without it they still use whichever authorization server the MCP server names.
  • Credentials saved without `issuer`: tokens and client information your OAuthClientProvider persisted (file, keychain, database) before upgrading, including everything 1.x saved before 1.31.0. They still go to whichever authorization server is named at first use. Add issuer to them, or clear them so users sign in again.
  • Your own `OAuthClientProvider`: save exactly what saveTokens() and saveClientInformation() are given, including issuer. For pre-registered credentials, include issuer in what clientInformation() returns.

Not covered by this fix:

  • a new interactive sign-in, which still goes to the authorization server the MCP server names. Only complete sign-ins for servers you trust.
  • refreshAuthorization() and exchangeAuthorization() called directly
  • 2.x with skipIssuerMetadataValidation: true

If an affected client may have connected to an untrusted MCP server, rotate its client secret or signing key and revoke its tokens.

If you cannot upgrade yet, connect OAuth clients only to MCP servers you trust. 2.0.0 and 2.1.0 already accept expectedIssuer.

Credited to Aviral2642, AlexMelanFromRingo, Gal3m, JosephDoUrden, Igfray, OriginalKazdov, lwebmedia.

Questions

Is upgrading to 1.31.0 or 2.2.0 enough?

Not always. Bundled providers need expectedIssuer, and credentials saved before the upgrade need an issuer added or must be cleared. Custom providers must persist issuer.

Does a stdio MCP client have this problem?

No. The advisory says stdio clients and MCP servers built with the SDK are not affected. The flaw is in the HTTP OAuth client.

Do I need to rotate secrets?

Yes, if the client may have connected to an untrusted MCP server. Rotate the client secret or signing key and revoke the tokens, since a hostile server could already have received them.

References

How this page is made: the data panel comes from the GitHub Advisory Database, CISA’s KEV catalog, the NVD and FIRST EPSS, and updates twice a day. The analysis is written by an AI model from the sources listed above and checked by automated rules before it is published. Check versions against the vendor's advisory before you act. Methodology

Coverage on the Beam blog

More MCP TypeScript SDK vulnerabilities

All MCP TypeScript SDK entries