CVE-2026-54311: n8n: prototype pollution leaks cross-user workflow data

## Impact An authenticated user with permission to create or modify workflows could pollute the sandbox used by the Merge node's SQL Query mode. Because the sandbox context was cached and reused across all workflow executions on the instance, prototype mutations introduced by one user's workflow persist into subsequent Merge SQL executions belonging to other users or projects. This allowed a low-privileged attacker to intercept workflow data processed by other users on the same instance. This issue only affects multi-user n8n instances where more than one user has permission to create and execute workflows containing the Merge node in SQL Query mode. ## Patches The issue has been fixed in n8n versions 2.25.7, and 2.26.2. Users should upgrade to one of these versions or later to remediate the vulnerability. ## Workarounds If upgrading is not immediately possible, administrators should consider the following temporary mitigations: - Limit workflow creation and editing permissions to fully trusted users only. - Disable the Merge node by adding `n8n-nodes-base.merge` to the `NODES_EXCLUDE` environment variable. These workarounds do not fully remediate the risk and should only be used as short-term mitigation measures.

A low-privileged attacker — for example, a contractor with workflow creation rights on a shared enterprise n8n instance used for AI agent automation — creates a workflow containing a Merge node configured in SQL Query mode. The SQL payload is crafted to mutate JavaScript prototypes within the shared sandbox (e.g., overriding Object.prototype properties to redirect property accesses). Because n8n caches and reuses this sandbox context across all workflow executions on the instance, the prototype mutations persist after the attacker's workflow runs. When other users subsequently execute their own workflows containing Merge nodes — such as a security analyst running an automated CVE triage pipeline or an ops engineer processing customer data through an LLM — the corrupted prototype chain causes their workflow data to be intercepted or surfaced in attacker-accessible execution context. The attacker reviews captured payloads to extract API credentials, LLM outputs, or sensitive business records, all without any interaction from victim users beyond their routine workflow execution.