NewApp Monitoring is liveSign up free
AgentbeamBeta
AI agent vulnerability · Other MCP servers and tools

GHSA-8g7g-hmwm-6rv2: n8n-mcp affected by path traversal, redirect-following SSRF, and telemetry payload exposure

highCVSS 3.1 8.3

n8n-mcp before 2.50.1 has three flaws in deployments that use the n8n API integration: unvalidated workflow identifiers used in URL paths, webhook URLs that follow redirects (a non-blind SSRF), and telemetry that uploads unredacted workflow edits, which can include tokens and API keys. An authenticated MCP caller can trigger the first two. Fixed in 2.50.1.

Published May 8, 2026 · Analysis Oct 28, 2026

What an attacker can do

n8n-mcp lets AI assistants manage n8n workflows over MCP. Three independent problems were reported.

First, the n8n API client used caller-supplied identifiers as URL path segments without validation. An authenticated MCP caller who passes a crafted workflow id can steer outbound requests, carrying the configured n8n API key, to other same-origin endpoints. That bypasses handler-level access controls, including the DISABLED_TOOLS setting.

Second, webhook, form and chat trigger URLs that passed validation still followed redirects. A URL could redirect the outbound request to a host that validation would have rejected, and the response body goes back to the caller, so this is non-blind SSRF.

Third, mutation telemetry, which is opt-in by default, uploaded partial-update operation diffs without redaction. Those values can contain the node parameters of the workflow: bearer tokens, API keys and webhook secrets. Whoever receives or can read that telemetry may see them.

The attacker needs an authenticated MCP caller and an n8n API integration with an API key. That includes a prompt-injected agent that calls the tools. The rating is CVSS 8.3, with high confidentiality and integrity impact because the n8n API key is used in the redirected or traversed requests.

Who is affected

Affected: the npm package n8n-mcp before 2.50.1. Exposure needs the n8n API integration (an n8n API key configured). The first two issues are reachable through the HTTP transport. The advisory says stdio mode exposes no HTTP surface for them. The telemetry issue applies to instances where telemetry is on.

Check the version:

npm ls n8n-mcp

If you start it with npx or Docker, check the version npx resolves or the image tag.

n8n-mcp

Ecosystem
npm
Affected
< 2.50.1
Fixed in
2.50.1
Check
npm ls n8n-mcp

How to fix it

Upgrade to n8n-mcp 2.50.1 or later:

npm install [email protected]

Workarounds from the advisory:

  • For the path and redirect issues, restrict network access to the HTTP transport with a firewall, reverse-proxy ACL or VPN so only trusted callers reach the MCP port, or switch to stdio mode.
  • For the telemetry issue, set N8N_MCP_TELEMETRY_DISABLED=true before starting the server, or run npx n8n-mcp telemetry disable once.

If telemetry was on while workflows held secrets, treat those tokens, API keys and webhook secrets as exposed and rotate them.

How to tell if you were hit

Check whether telemetry was enabled on the instance, and whether any workflow you edited through MCP held credentials in node parameters. If so, rotate them.

In MCP and n8n API logs, look for workflow ids containing slashes, dots or encoded path characters, and for requests to n8n API paths the calling tool should not reach. For redirects, look for trigger-webhook calls whose target answered with a 3xx and was followed to an internal or unexpected host. Review calls to tools that you disabled with DISABLED_TOOLS that still ran.

Original advisory

From the GitHub Advisory Database (CC BY 4.0).

Impact

n8n-mcp versions before 2.50.1 contained three independently-reported issues affecting deployments that run the n8n API integration:

  1. Caller-supplied identifiers were not validated before being used as URL path segments by the n8n API client. An authenticated MCP caller passing a crafted workflow id could cause outbound requests carrying the configured n8n API key to land on other same-origin endpoints, bypassing handler-level access controls (including DISABLED_TOOLS).
  1. Validated webhook, form, and chat trigger URLs followed redirects. A URL that passed initial validation could redirect the outbound request to a host that would otherwise have been rejected, with the response body returned to the caller. Reachable as non-blind SSRF over authenticated MCP calls.
  1. Mutation telemetry stored unredacted operation payloads. On instances running with the default opt-in telemetry, partial-update operation diffs were uploaded without redaction. Operation values can carry the same node-parameter values the workflow contains, including bearer tokens, API keys, and webhook secrets.

Severity

CVSS 8.3 (HIGH). Exploitation requires an authenticated MCP caller and an n8n API integration configured with an n8n API key.

Patched versions

Upgrade to n8n-mcp >= 2.50.1.

Workarounds

  • For issues (1) and (2): restrict network access to the HTTP transport (firewall, reverse-proxy ACL, or VPN) so only trusted callers can reach the MCP HTTP port; or switch to stdio mode, which exposes no HTTP surface for these issues.
  • For issue (3): set N8N_MCP_TELEMETRY_DISABLED=true in the environment before starting the server, or run npx n8n-mcp telemetry disable once.

Credit

Reported by @cybercraftsolutionsllc.

Credited to cybercraftsolutionsllc.

Questions

Does exploitation need authentication?

Yes. The caller must be an authenticated MCP caller, and the n8n API integration must be configured with an API key.

Are stdio deployments affected?

The advisory says stdio mode exposes no HTTP surface for the path and redirect issues, so switching to it is a listed workaround. The telemetry leak is separate.

How do I turn off the telemetry leak without upgrading?

Set N8N_MCP_TELEMETRY_DISABLED=true before starting the server, or run npx n8n-mcp telemetry disable once.

Is this exploited in the wild?

No source reports exploitation. It is not in CISA's KEV catalog.

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

More Other MCP servers and tools vulnerabilities

All Other MCP servers and tools entries