CVE-2025-57760
8.8Langflow · Langflow
A privilege escalation vulnerability in Langflow allows authenticated users with Remote Code Execution capabilities to create administrative superuser accounts.
Executive summary
A critical privilege escalation vulnerability in Langflow allows authenticated users to gain full administrative control, posing a significant risk to system integrity.
Vulnerability
This vulnerability, categorized as Improper Privilege Management (CWE-269), occurs when an authenticated user leverages existing Remote Code Execution access to invoke the internal CLI command langflow superuser. This action permits the creation of a new administrative account, granting full superuser privileges regardless of the user's original account status.
Business impact
Successful exploitation of this vulnerability allows an attacker to achieve full administrative control over the Langflow instance. This can lead to total compromise of sensitive AI workflows, unauthorized data access, and potential persistence within the infrastructure. Given the CVSS score of 8.8, this flaw represents a high-severity risk that could facilitate lateral movement or data exfiltration within the organization.
Remediation
Immediate Action: Update Langflow to version 1.5.1 and langflow-base to version 0.5.1 immediately to resolve the privilege escalation flaw.
Proactive Monitoring: Monitor system logs for unauthorized use of CLI commands or the creation of unexpected administrative user accounts.
Compensating Controls: Restrict network access to the Langflow instance and enforce strict access controls on the underlying container environment to prevent the initial Remote Code Execution required to trigger this flaw.
Exploitation status
Public Exploit Available: Unknown
Analyst recommendation
The severity of this vulnerability necessitates immediate action. Administrators must prioritize the deployment of the patched versions identified by the vendor to eliminate the privilege escalation vector. Failure to patch allows any user with minor execution capabilities to escalate to full administrative rights, effectively granting them complete control over the platform and its hosted AI workflows.