Executive brief
n8n-mcp is a server that allows AI assistants to interact with n8n automation workflows. A security flaw in versions prior to 2.47.13 causes the server to record sensitive information, such as API keys and login tokens, into its internal activity logs when running in HTTP mode. If these logs are shared with other systems or viewed by unauthorized personnel, it could lead to the theft of credentials and unauthorized access to connected services.
Technical details
n8n-mcp prior to version 2.47.13 is vulnerable to sensitive information disclosure (CWE-532) when operating in HTTP transport mode. The request dispatcher and related code paths write full JSON-RPC parameters and tool arguments to server logs before any redaction occurs. This includes sensitive data such as bearer tokens, OAuth credentials, and per-tenant API keys passed via 'n8n_manage_credentials'. An authenticated attacker or a legitimate user performing standard operations can trigger the logging of these secrets. The vulnerability is mitigated if using the 'stdio' transport mode, but in HTTP mode, the existing console-silencing layer is insufficient to prevent the data from being passed to the logger. The issue is fixed in version 2.47.13 by implementing a metadata-only summarizer for tool-call arguments.
Affected products
- czlonkowski n8n-mcp <= 2.47.12
Timeline
- 2026-04-20: patched: Version 2.47.13 released
- 2026-04-21: advisory: GitHub Security Advisory published
- 2026-05-08: disclosed: CVE published to NVD