Critical Langflow Vulnerability Exploited to Execute Malicious Code

A critical security vulnerability in Langflow, tracked as CVE-2026-5027, is being actively exploited in the wild, allowing attackers to execute arbitrary code on affected systems. The flaw resides in the POST /api/v2/files endpoint, where the filename parameter from multipart form data is not properly sanitized. Attackers can inject path-traversal sequences such as ../ to write files to arbitrary locations on the server filesystem, ultimately achieving remote code execution.

The vulnerability carries a CVSS v3 score of 8.8, indicating high severity. It requires low complexity, minimal privileges, and no user interaction, making it particularly dangerous for internet-facing deployments. According to Tenable's advisory (TRA-2026-26), the issue was discovered and reported by researcher Joshua Martinelle. Despite multiple disclosure attempts starting January 20, 2026, the vendor did not respond within the expected timeframe, leading to public disclosure on March 27, 2026.

Threat intelligence and exploit tracking teams have flagged the issue as a high-priority risk. Early exploitation signals, including discussions on LinkedIn, indicate that attackers are actively scanning for and exploiting the Langflow path traversal flaw. As proof-of-concept code spreads within both security and attacker communities, opportunistic scanning and automated exploitation are expected to increase.

Langflow is an open-source tool used for building and deploying AI workflows, particularly those involving large language models (LLMs). Its popularity in development and production environments makes it an attractive target. Organizations that expose Langflow instances to the internet or integrate it into critical pipelines are at heightened risk.

At the time of disclosure, no official patch or fix had been released. This significantly increases the risk for organizations using Langflow. Security teams are advised to implement temporary mitigations such as restricting access to the vulnerable endpoint, applying strict input validation controls, and monitoring systems for suspicious file activity. The lack of a vendor response also raises concerns about patch management and coordinated disclosure practices.

Given the nature of the vulnerability, attackers could potentially chain this flaw with other weaknesses to escalate privileges or establish persistence within compromised systems. Organizations should prioritize threat hunting and log analysis to detect any signs of exploitation attempts. Users of Langflow are strongly encouraged to review their deployments, limit exposure, and stay up to date with any future security patches or advisories from the vendor.

This incident serves as a stark reminder of the importance of secure coding practices, particularly around file handling and input validation. As exploitation risks continue to grow, the Langflow case highlights the dangers of delayed remediation in widely used tools and the critical need for proactive vulnerability management.