Threat Intelligence Database

n8n versions before 1.123.64 do not properly mask custom HTTP header credentials in LLM sub-node execution data. This results in plaintext API keys and secrets being written to workflow execution records. Authenticated users with access to these execution records can read and export these exposed credentials, which persist in the database. This vulnerability is categorized under CWE-532 (Exposure of Sensitive Information). The advisory is a duplicate and has been withdrawn in favor of GHSA-89gh-3pgc-v5h2.

n8n's JavaScript task runner shared a module cache across all users executing Code nodes, allowing a user to poison this cache and affect other users' Code-node executions on the same runner. This vulnerability breaks user isolation within a single n8n instance but does not enable sandbox escape or remote code execution. It affects all multi-user n8n instances running the JS task runner with built-in or external modules enabled. The issue is fixed in n8n versions 1.123.67, 2.31.5, and 2.32.1. Temporary mitigations include restricting access to trusted users, disabling module access in Code nodes, or using dedicated external runners per user or project.

A SQL injection vulnerability exists in the Postgres Trigger node of n8n, allowing authenticated users to inject arbitrary SQL commands. This occurs because user-supplied identifier parameters are interpolated into SQL statements without proper escaping. Exploitation grants full read and write access to the connected PostgreSQL database using the configured credentials. The vulnerability affects n8n versions prior to 1.123.67 and versions from 2.32.0 up to but not including 2.32.1. Fixed versions are 1.123.67, 2.31.5, and 2.32.1 or later. Temporary mitigations include restricting access to trusted users, disabling the PostgresTrigger node, and limiting PostgreSQL credential privileges. These mitigations do not fully remediate the risk and upgrading is recommended.

The n8n Snowflake node's Execute Query operation was vulnerable to SQL injection due to direct interpolation of expression values into SQL strings without parameterization. This vulnerability affects workflows that embed untrusted expression data directly in raw SQL queries. The issue is fixed in n8n versions 1.123.67, 2.31.5, and 2.32.1 by introducing a parameter binding feature. Temporary mitigations include restricting workflow editing to trusted users and auditing existing workflows for unsafe query usage.

n8n versions before 2.30.1 contain a privilege escalation vulnerability in the AI Agents feature. The vulnerability arises because the node-execution tool does not enforce proper authorization checks. As a result, a Project Viewer user can escalate privileges by interacting with an agent that has node tools enabled, allowing execution of arbitrary nodes and unauthorized access to credential secrets. This advisory is a duplicate and has been withdrawn in favor of GHSA-x5vx-c2c8-m3w9.

A race condition vulnerability exists in the Git clone node of n8n before versions 1.123.64, 2.29.8, and 2.30.1. Authenticated users can exploit a time-of-check to time-of-use (TOCTOU) issue by replacing a directory with a symlink after path validation but before cloning. This allows planting a crafted repository that n8n loads as a custom node on restart, leading to arbitrary JavaScript execution on the server. Both self-hosted and cloud instances are affected. This advisory is a duplicate and has been withdrawn in favor of GHSA-g3r5-9h93-4j2c.

n8n versions before 1.123.64, 2.x before 2.29.8, and before 2.30.1 contain a DOM-based cross-site scripting (XSS) vulnerability in the HTML preview feature. The vulnerability arises because the execution output is rendered into an iframe srcdoc without the sandbox attribute, allowing a sanitizer bypass that enables injected scripts to execute with the same origin as the editor. An attacker with global:member privileges can exploit this by causing a victim to open the preview, enabling the injected script to call authenticated APIs using the victim's session.

n8n before versions 1.123.64, 2.29.8, and 2.30.1 contains a stored DOM cross-site scripting (XSS) vulnerability in the Resource Locator component. The vulnerability occurs because the cachedResultUrl parameter is passed to window.open() without validating the URL scheme. An attacker with workflow creation or editing privileges can embed a javascript: scheme in cachedResultUrl, causing malicious scripts to execute in the browser of a user who opens the crafted workflow.

n8n versions before 1.123.64, 2.29.8, and 2.30.1 have a vulnerability where the full Google Service Account private key is exposed in the JWT header's kid field. This exposure occurs because the private key was mistakenly placed in a Base64-encoded JWT header, which is not encrypted. An attacker with access to the JWT could recover the private key and impersonate the service account, potentially accessing or modifying authorized Google Cloud resources. Only instances configured with Google Service Account credentials are affected. This advisory is a duplicate and has been withdrawn in favor of GHSA-9r8p-h6cc-6qhm.

n8n versions before 2.30.1 and 2.29.8 have a vulnerability where all Public API key scopes are assigned to JWTs issued via the Token Exchange module regardless of the user's role. This allows a low-privileged user with a valid external JWT to perform administrator-only Public API actions such as role escalation, user creation, and deletion. If unverified Community Package installation is enabled, remote code execution is possible. This advisory is a duplicate and has been withdrawn in favor of GHSA-777w-rpr6-c52h.