CVE-2026-72774: 
n8n vulnerability analysis and mitigation

CVE-2026-72774 is a credential authorization bypass vulnerability in the HTTP Request node of n8n, an open-source workflow automation platform. An authenticated member with edit access to a shared workflow can reference and exfiltrate another user's credential by specifying the credential type via an expression, bypassing ownership checks. The vulnerability affects n8n versions before 1.123.67 (v1 branch), 2.31.5 (v2.31 branch), and 2.32.1 (v2.32 branch). It was published on August 11, 2026, with the security advisory authored by csuermann and credited to researcher g4nkd. It carries a CVSS v4.0 base score of 7.1 (High) (GitHub Advisory, GitHub Advisory DB).

The root cause is classified as CWE-639 (Authorization Bypass Through User-Controlled Key): the pre-execution permission check in the HTTP Request node evaluates the genericAuthType field against the unresolved expression string rather than the resolved credential type at runtime. Because the ownership check compares the raw expression (e.g., ={{ $json.credType }}) instead of the actual credential type it resolves to, the check is effectively skipped, and the credential is loaded and used at execution time. Exploitation requires the attacker to be an authenticated n8n member with edit access to a shared workflow and to know the numeric or string identifier of the target credential. No public proof-of-concept code has been identified (GitHub Advisory, GitHub Advisory DB).

Successful exploitation allows an authenticated but lower-privileged n8n user to access, use, or exfiltrate credentials belonging to other users that they were never explicitly granted access to. This primarily affects confidentiality — stored credentials (e.g., API keys, OAuth tokens, passwords) can be extracted and used externally, potentially enabling lateral movement into third-party systems integrated with the n8n instance. Integrity of subsequent systems is also at risk, as the attacker can use the stolen credentials to perform unauthorized actions in connected services (GitHub Advisory).

There is no evidence of active in-the-wild exploitation or a publicly available proof-of-concept exploit as of the time of publication. The NVD SSVC assessment confirms exploitation status as "none" and the attack is not automatable, requiring the attacker to know the target credential's identifier. The EPSS score is approximately 0.295% (22nd percentile), indicating a low near-term exploitation probability. The vulnerability is not listed in the CISA Known Exploited Vulnerabilities (KEV) catalog (GitHub Advisory DB, VulnCheck Advisory).

1. Reconnaissance: As an authenticated n8n member, identify shared workflows to which you have edit access. Browse the n8n UI or API to enumerate workflow configurations and look for credential identifiers (numeric IDs) used by other users in the instance.
2. Identify target credential ID: Attempt to discover the target credential's identifier through workflow metadata, API responses (e.g., GET /rest/credentials), or social engineering. The credential ID is required for exploitation.
3. Craft malicious workflow node: In an HTTP Request node within the shared workflow, set the genericAuthType field to an expression (e.g., ={{ 'predefinedCredentialType' }}) that resolves to the desired credential type at runtime, while referencing the target credential's ID in the credential field.
4. Trigger workflow execution: Execute the workflow. Because the pre-execution permission check evaluates the unresolved expression string rather than the resolved credential type, the ownership check is bypassed and the target credential is loaded.
5. Exfiltrate credential: Configure the HTTP Request node to send the request to an attacker-controlled server, causing the credential (e.g., API key, bearer token) to be transmitted in the request headers or body, completing the exfiltration (GitHub Advisory).

• Logs: n8n execution logs showing HTTP Request node executions referencing credential IDs not owned by the executing user; workflow execution records where genericAuthType is set to an expression value rather than a static string.
• Network: Outbound HTTP requests from the n8n server to unexpected or attacker-controlled external endpoints, particularly those carrying authentication headers (e.g., Authorization: Bearer ..., X-API-Key) associated with credentials belonging to other users.
• Application Behavior: Credential usage audit logs (if enabled) showing a credential being used in a workflow execution by a user who is not the credential owner and was not explicitly granted access; repeated workflow executions by a non-owner user against workflows containing HTTP Request nodes with expression-based genericAuthType fields.

n8n has released patched versions: 1.123.67 (v1 branch), 2.31.5 (v2.31 branch), and 2.32.1 (v2.32 branch). Upgrading to one of these versions or later is the recommended remediation. If immediate upgrade is not possible, administrators should: (1) restrict n8n instance access to fully trusted users only; (2) exclude the HTTP Request node by adding n8n-nodes-base.httpRequest to the NODES_EXCLUDE environment variable if the node is not required; and (3) audit credential sharing and workflow access to limit exposure of credential IDs to untrusted users. These workarounds do not fully remediate the risk and should only be used as short-term measures (GitHub Advisory).