CISA orders agencies to patch critical Langflow flaw

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has issued a directive requiring all federal agencies to patch a critical vulnerability in the Langflow visual framework for building AI agents by Friday. The actively exploited flaw poses a significant risk to national security, prompting immediate action from the agency.

The vulnerability, tracked as CVE-2025-XXXX, allows attackers to execute arbitrary code on systems using Langflow, a popular open-source tool for designing and deploying AI workflows. CISA added the flaw to its Known Exploited Vulnerabilities (KEV) catalog, signaling that it is being actively targeted by malicious actors. Federal civilian agencies must apply the patch within the week to comply with Binding Operational Directive (BOD) 22-01, which mandates timely remediation of known exploited vulnerabilities.

Langflow, developed by Logspace, is widely used for building AI agents and chatbots without extensive coding. The vulnerability affects versions before 1.1.0, and users are urged to upgrade immediately. While CISA’s directive applies specifically to federal agencies, the agency strongly recommends that private sector organizations also apply the patch to prevent potential breaches.

Security researchers have observed active exploitation attempts in the wild, though CISA has not disclosed specific details about the attacks or the threat actors involved. The agency’s move underscores the growing urgency to secure AI development tools as their adoption accelerates across industries.