Hackers Start Exploiting Critical Langflow Vulnerability

CVE-2026-0768 is a critical-severity remote code execution vulnerability in Langflow's custom component editor code validator. The vulnerability arises because user-supplied strings are not properly validated before being executed as Python code, enabling unauthenticated attackers to run arbitrary code with root privileges. The flaw affects all Langflow releases up to version 1.4.2. It was reported through ZDI in July 2025 and publicly disclosed in January 2026. Since then, threat actors have exploited this vulnerability to perform reconnaissance and credential harvesting, with over 360 exploitation attempts observed hitting canaries in the UK and significant activity traced to Russia. The CVSS score is 9.8, indicating critical severity.

Successful exploitation allows unauthenticated remote attackers to execute arbitrary Python code as root on vulnerable Langflow instances. This can lead to full system compromise, including unauthorized access to environment variables, secret keys, SSH credentials, and other sensitive data. Observed exploitation activity includes reconnaissance and credential harvesting, increasing the risk of further attacks and lateral movement within affected environments.

A fix is available in Langflow versions after 1.4.2. Users should upgrade to a patched version beyond 1.4.2 to remediate this vulnerability. Until patched, restricting access to Langflow instances and monitoring for suspicious activity related to this vulnerability is advised. Check the official Langflow vendor advisory or trusted sources for the latest patch and remediation guidance.