/

CVE-2026-72768

Severity:

HIGH RISK

(8.3)

SSRF Protection Bypass (CWE-918) in n8n MCP Client

CVE-2026-72768

Analyzed

Vulnerability type

Category:

Server-Side Request Forgery

CWE:

CWE-918

Impact:

Unauthorized access to internal services and sensitive response data

Severity:

HIGH RISK

(8.3)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

A server-side request forgery protection bypass in the n8n MCP Client node allows authenticated users to create workflows that avoid intended SSRF controls.

Under conditions where an attacker can create or run workflows, requests may reach internal or blocked hosts, with response data returned through the workflow.

Affected context

The vulnerability occurs in n8n deployments that expose the MCP Client node to authenticated users who can create or execute workflows. Requests generated by this node may bypass the platform's SSRF protection path, allowing access to destinations that would otherwise be restricted, depending on network reachability and workflow permissions.

Exploit preconditions

  • Attack surface: The n8n MCP Client node and its outbound request handling
  • Required access: Authenticated access with permission to create or execute workflows using the affected node
  • Trigger point: Execution of a crafted workflow containing an MCP Client request
  • Why exploitation is possible: The node can send requests without consistently passing them through the intended SSRF protection controls

Why this happens

The underlying issue is an incomplete enforcement of destination validation for outbound requests generated by the MCP Client node. A request path can reach network resources without applying the platform's SSRF restrictions, allowing user-controlled workflow behavior to influence destinations that should be blocked.

This maps to CWE-918: Server-Side Request Forgery, where a server-side component is induced to make requests to attacker-selected internal, external, or otherwise restricted resources.

Potential impact

  • An attacker may access internal administrative or cloud metadata services
  • Sensitive response data could be read through workflow outputs
  • Internal network topology or service information may be disclosed
  • Compromised internal services could be targeted for further abuse
  • The n8n instance or connected systems could face increased lateral-movement risk

General mitigation guidance

Mitigation generally focuses on enforcing SSRF controls consistently across every outbound request path, including MCP Client operations.

Common defensive patterns include:

  • Routing all user-influenced outbound requests through a centralized, policy-enforcing request broker
  • Applying destination validation, scheme restrictions, DNS resolution checks, and private-network blocking before connection establishment
  • Revalidating destinations after redirects and preventing alternate address representations from bypassing policy
  • Restricting MCP Client availability to trusted roles and workflows that require it
  • Limiting workflow execution privileges and isolating n8n from sensitive network segments
  • Using network egress controls as a defense-in-depth measure
  • Returning only the minimum response data required by the workflow

How teams detect issues like this

Detection approaches can include:

  • Static analysis and code review for outbound request paths that do not use centralized SSRF validation
  • Tests that verify every supported protocol, redirect behavior, hostname form, and resolved address is checked
  • Review of workflow permissions and use of MCP Client nodes by authenticated users
  • Network monitoring for unexpected connections from n8n workers to private address ranges, metadata endpoints, or restricted services
  • Policy enforcement requiring approved egress gateways for application-initiated requests
  • Automated security review tools that flag user-controlled URLs and inconsistent destination validation

Sources

SourceLink
github.comgithub.com/n8n-io/n8n/security/advisories/GHSA-vhf8-cg2h-cg3p
vulncheck.comvulncheck.com/advisories/n8n-before-ssrf-protection-bypass-via-mcp-client

Frequently asked

questions

Exploitation generally requires authentication plus permission to create or execute workflows that use the affected MCP Client node.

Not necessarily. Exposure depends on n8n's network connectivity, outbound filtering, destination controls, service authentication, and the permissions available to the attacker.

Ensure every MCP Client outbound request is processed by centralized SSRF protections that validate destinations before connection and after redirects or DNS resolution.

Combine strict workflow permissions with deny-by-default egress controls, comprehensive request-path testing, private-network isolation, and monitoring for unexpected outbound connections.

Find out what's already
exploitable in your codebase.