Tools AI Risk Radar ai-incident-0153
Incident record
MCP SDK fixes credentials sent to untrusted authentication servers
- Severity
- High
- Status
- Patched
- Type
- Infra Vuln
- Target
- @modelcontextprotocol/sdk, @modelcontextprotocol/client
- Actor
- researcher
- CVE
- CVE-2026-104850
What happened
A malicious MCP server could direct an HTTP OAuth client to an attacker-controlled authentication server. A client already holding credentials for another issuer could send its refresh token or client secret to that destination.
@modelcontextprotocol/sdk: affected >= 1.12.0, < 1.31.0; fixed in 1.31.0. @modelcontextprotocol/client: affected >= 2.0.0, < 2.2.0; fixed in 2.2.0. The date shown is the maintainer’s original disclosure date. Its advisory assigns 7.5 under CVSS 3.1. Applications using the affected provider patterns must also configure the expected issuer or migrate their custom provider as the advisory directs.
Sources
- Primary disclosuregithub.com/modelcontextprotocol/typescript-sdk/security/advi…
One record from the AI Risk Radar, maintained by the Security Desk.Data: CSV · JSON ·RSS · CC BY 4.0 with attribution to YFarmX.Tracker updated · 13 October 2026