1,423 Attacks Target Critical Langflow Flaw Allowing Non-Admin Code Execution

Security researchers at Previdian have detected over 1,400 exploitation attempts targeting a critical remote code execution vulnerability in Langflow, the open-source AI workflow platform. The flaw, tracked as CVE-2026-105740 with a CVSS score of 9.9, allows any authenticated user—including those with non-administrator privileges—to execute arbitrary operating system commands on the server by adding a malicious Model Context Protocol (MCP) server configuration.