Does MCP SDKs and registry have known security vulnerabilities?

Yes. The Injection has tracked 1 story about MCP SDKs and registry with a published CVE: CVE-2026-104850 (CVSS 7.5), OAuth client could leak credentials to servers. Check each story for fixed versions.

What is the latest MCP SDKs and registry security issue?

The most recent story is "MCP TypeScript SDK: OAuth client could leak credentials to servers" (high severity), disclosed September 30, 2026 by Model Context Protocol. CVE-2026-104850 in the MCP TypeScript SDK let a malicious MCP server pick the authorization server that receives a client's OAuth refresh tokens and client secrets. Fixed in @modelcontextprotocol/sdk 1.31.0 and client 2.2.0.

Which attack techniques have been used against MCP SDKs and registry?

Stories about MCP SDKs and registry on The Injection most often involve mcp (model context protocol), credential theft, ai threat intelligence, prompt injection. Each technique has a glossary entry with a definition and every related story.

MCP SDKs and registry security

The official Model Context Protocol SDKs, clients and server registry.

The Injection has tracked 3 security stories affecting MCP SDKs and registry since August 27, 2026 (vulnerability 1, research 1, supply chain 1), 2 rated high or critical. The most recent, disclosed September 30, 2026, is "MCP TypeScript SDK: OAuth client could leak credentials to servers".

Stories

High or critical

With a CVE

2026-09-30

Latest disclosure

CVEs

CVE	CVSS	Severity	Story
CVE-2026-104850	7.5	High	MCP TypeScript SDK: OAuth client could leak credentials to servers

Every MCP SDKs and registry story

2026-09-30HIGH
Vulnerability · Model Context Protocol · CVE-2026-104850 · CVSS 7.5
MCP TypeScript SDK: OAuth client could leak credentials to servers
CVE-2026-104850 in the MCP TypeScript SDK let a malicious MCP server pick the authorization server that receives a client's OAuth refresh tokens and client secrets. Fixed in @modelcontextprotocol/sdk 1.31.0 and client 2.2.0.
2026-09-30INFO
Research · Google Threat Intelligence Group
Google GTIG: half of 2026 AI flaws hit agent orchestration
Google Threat Intelligence Group tracked 2,076 vulnerabilities in AI software from January 2025 to August 2026, more than 1,500 of them in 2026, and says agent orchestration frameworks account for half of them.
2026-08-27HIGH
Supply chain · AIR Security
MCPJacking: 155 hijackable servers in the official MCP registry
MCPJacking is an attack on MCP registry entries whose domains have expired. AIR Security found 155 such servers in the official MCP registry, re-registered the domains and gained remote prompt execution on agents that trusted them.

Questions people ask

Does MCP SDKs and registry have known security vulnerabilities?: Yes. The Injection has tracked 1 story about MCP SDKs and registry with a published CVE: CVE-2026-104850 (CVSS 7.5), OAuth client could leak credentials to servers. Check each story for fixed versions.
What is the latest MCP SDKs and registry security issue?: The most recent story is "MCP TypeScript SDK: OAuth client could leak credentials to servers" (high severity), disclosed September 30, 2026 by Model Context Protocol. CVE-2026-104850 in the MCP TypeScript SDK let a malicious MCP server pick the authorization server that receives a client's OAuth refresh tokens and client secrets. Fixed in @modelcontextprotocol/sdk 1.31.0 and client 2.2.0.
Which attack techniques have been used against MCP SDKs and registry?: Stories about MCP SDKs and registry on The Injection most often involve mcp (model context protocol), credential theft, ai threat intelligence, prompt injection. Each technique has a glossary entry with a definition and every related story.