Flowise SSRF Bypass: DNS Rebinding Opens LLM Flows to Attackers

The National Vulnerability Database has disclosed CVE-2026-41272, a high-severity Server-Side Request Forgery (SSRF) vulnerability in Flowise, a drag-and-drop interface for building custom large language model (LLM) flows. Prior to version 3.1.0, Flowise’s core security wrappers, secureAxiosRequest and secureFetch, designed to prevent SSRF, contain critical logic flaws. These issues allow attackers to bypass intended allow/deny lists.

The bypass mechanisms are twofold: attackers can exploit DNS Rebinding (Time-of-Check Time-of-Use) to trick the system into connecting to unauthorized internal resources, or leverage the default configuration, which fails to enforce any deny list at all. This means if your Flowise instance isn’t explicitly configured with a deny list, it’s wide open to SSRF.

This isn’t theoretical; SSRF can lead to internal network reconnaissance, access to metadata services, and even arbitrary code execution in some configurations. For any organization using Flowise, this vulnerability presents a direct path for attackers to pivot from an exposed LLM interface into your internal infrastructure. Patching to version 3.1.0 is non-negotiable.