UTC
High

mcp-handler has a tool response leak across concurrent client sessions ('Race Condition')

GHSA-w2fm-25vw-vh7f · published 2026-04-01

Severity
High
CVSS score
7.1
Chance of exploitation (EPSS)
No CVE
Fixed in
1.1.0
What it is

A low-privileged attacker making concurrent requests to an mcp-handler endpoint can read another client's session data, including authentication information and tool execution state. This is a confidentiality breach with potential for limited integrity impact. Root Cause: CVE-2026-25536 in @modelcontextprotocol/sdk < 1.26.0. The SDK did not prevent reuse of stateless transports across client connections.

What to do

Upgrade mcp-handler to 1.1.0 or later. If you start it from an MCP config with npx or uvx, pin that version so an old cached copy is not used.

Upgrade to [email protected]. This release raises the minimum peer dependency to @modelcontextprotocol/sdk@>=1.26.0, which contains the fix for CVE-2026-25536.

If you cannot upgrade yet

Upgrade @modelcontextprotocol/sdk to >=1.26.0 (note: the SDK will throw on transport reuse, which will break mcp-handler < 1.1.0 which effectively forces the upgrade) Alternatively, manually create fresh McpServer and transport instances per request in your handler code

Affected
PackageRegistryAffectedFixed in
mcp-handlernpm< 1.1.01.1.0
Downstream in the Official MCP Registry

7 registered MCP servers install an affected version today, by the dependency versions a fresh install resolves.

  • io.github.meltingpixelsai/harvey-budgetmcp-handler@1.0.2
  • io.github.meltingpixelsai/harvey-intelmcp-handler@1.0.2
  • io.github.meltingpixelsai/harvey-verifymcp-handler@1.0.2
  • io.github.meltingpixelsai/zero-core-budgetmcp-handler@1.0.2
  • io.github.meltingpixelsai/zero-core-intelmcp-handler@1.0.2
  • io.github.meltingpixelsai/zero-core-toolsmcp-handler@1.0.2
  • io.github.meltingpixelsai/zero-core-verifymcp-handler@1.0.2
See the trace →

The impact, patch and workaround text comes from the advisory, shortened (GitHub Advisory Database, CC BY 4.0). Proof-of-concept and reproduction sections and code are left out; the original advisory has them.

Sources: GitHub Advisory Database, reviewed and malware advisories (CC BY 4.0, https://github.com/advisories); advisories published by Cursor (https://github.com/cursor/cursor/security/advisories); advisories published by OpenClaw (https://github.com/openclaw/openclaw/security/advisories); advisories published by vLLM (https://github.com/vllm-project/vllm/security/advisories); advisories published by Open WebUI (https://github.com/open-webui/open-webui/security/advisories); advisories published by n8n (https://github.com/n8n-io/n8n/security/advisories); advisories published by Flowise (https://github.com/FlowiseAI/Flowise/security/advisories); advisories published by Langflow (https://github.com/langflow-ai/langflow/security/advisories); advisories published by the MCP Python SDK project (https://github.com/modelcontextprotocol/python-sdk/security/advisories); advisories published by MLflow (https://github.com/mlflow/mlflow/security/advisories); CISA Known Exploited Vulnerabilities catalog (CC0). Data as of 2026-11-22. ShieldMap describes how each problem works and how to fix it; it does not publish exploit code. Report a correction to [email protected].