UTC
Critical

Todoist MCP Server: Browser DNS rebinding lets websites write and delete Todoist data through local MCP server

CVE-2026-76807 · GHSA-833v-v58w-294h · published 2026-11-19

Severity
Critical
CVSS score
9.3
Chance of exploitation (EPSS)
0.22%
Fixed in
10.3.3
  • Exploitation Public proof of concept
  • Automatable No
  • Technical impact Total
  • CISA SSVC
What it is

A public website can operate a victim's local todoist-mcp-http server when the optional HTTP transport is running. The browser does not supply a Todoist API token. The MCP server supplies its configured TODOIST_API_KEY. The demonstrated impact is Todoist task creation and completion. The same browser-origin path exposes the rest of the registered Todoist write surface, including task updates, task rescheduling, project and section changes, comments, reminders, filters, labels, goals, assignments, and object deletion. This violates the browser-origin boundary. …

What to do

Upgrade @doist/todoist-mcp to 10.3.3 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.

Add strict Host validation before MCP dispatch. For local deployments, accept only explicit loopback hostnames such as 127.0.0.1, localhost, [::1], and any operator-configured trusted names. Reject browser requests with missing or untrusted Origin values before they reach StreamableHTTPServerTransport. Require a per-run local bearer token or equivalent caller secret for HTTP MCP mode, including loopback deployments. Host and Origin validation reduce DNS-rebinding exposure, but they are not caller authentication. Enable DNS rebinding protection in the MCP SDK HTTP transport if available, and pass explicit allowed host and origin lists. Keep stdio as the safest default local transport.

Affected
PackageRegistryAffectedFixed in
@doist/todoist-mcpnpm<= 10.3.210.3.3

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].