n8n privilege escalation in AI Agents via run_node_tool

Executive brief

n8n is a workflow automation platform that allows users to connect different applications. A security flaw in its AI Agents feature allows users with limited 'Project Viewer' permissions to bypass restrictions and perform actions they shouldn't be allowed to. This could lead to unauthorized access to sensitive credentials and the execution of arbitrary tasks within the system.

Technical details

A privilege escalation vulnerability exists in n8n's AI Agents feature due to missing authorization checks in the 'run_node_tool' component. An attacker with 'Project Viewer' privileges can exploit this by interacting with an AI agent that has node tools enabled. By crafting specific chat inputs, the attacker can trigger the execution of arbitrary nodes and retrieve sensitive credential secrets that should be restricted. This vulnerability is classified as Incorrect Authorization (CWE-863) and is addressed in version 2.30.1.

Affected products

• n8n-io n8n < 2.30.1

Timeline

• 2026-07-22: advisory: Initial advisory publication
• 2026-07-22: other: Advisory withdrawn as a duplicate of GHSA-x5vx-c2c8-m3w9

References

• https://api.github.com/users/Malayke
• https://github.com/Malayke
• https://api.github.com/users/Malayke/gists%7B/gist_id%7D
• https://api.github.com/users/Malayke/repos
• https://avatars.githubusercontent.com/u/4935500?v=4
• https://api.github.com/users/Malayke/events%7B/privacy%7D