Executive brief
Flowise is a low-code platform used to build and deploy custom AI applications and Large Language Model (LLM) workflows. A security flaw in several of its built-in tools allows attackers to bypass internal network protections and force the application to make unauthorized requests to internal servers. This could lead to the theft of sensitive cloud credentials, access to private internal data, or the compromise of other systems within the corporate network.
Technical details
Flowise implements a centralized HTTP security wrapper (httpSecurity.ts) designed to prevent Server-Side Request Forgery (SSRF) by enforcing IP deny-lists and pinning. However, several components—including the OpenAPI Toolkit, Web Scraper, MCP, and Arxiv tools—directly import and invoke raw HTTP clients like 'node-fetch' and 'axios' instead of using the secure wrapper. This architectural failure allows an attacker to bypass SSRF mitigations by providing malicious URLs to these specific tools. An attacker can exploit this to reach internal loopback addresses or cloud provider metadata services (e.g., 169.254.169.254) to exfiltrate IAM credentials or other sensitive environment data. The vulnerability is addressed in version 3.1.0 by refactoring these tools to use the secureFetch wrapper.
Affected products
- FlowiseAI Flowise < 3.1.0
- FlowiseAI flowise-components <= 3.0.13
Timeline
- 2026-04-15: advisory: GitHub Security Advisory published
- 2026-05-11: disclosed: CVE-2026-43995 published to NVD