Flowise, a drag & drop interface for building large language model flows, contains an endpoint /api/v1/public-chatbotConfig/:id that, before version 3.1.0, returns configuration data without any authentication. The response includes API keys, HTTP authorization headers, and credentials stored in password type fields, enabling an attacker who knows a chatflow UUID to acquire sensitive credentials and internal configuration, thereby compromising confidentiality and enabling credential theft.

The vulnerability affects FlowiseAI Flowise prior to version 3.1.0. All instances of the application running any version older than 3.1.0 and exposing the public‑chatbotConfig endpoint are impacted.

The CVSS score of 7.7 indicates a high severity vulnerability, yet the EPSS score of <1% suggests a low probability of exploitation. The issue is not listed in CISA’s KEV catalog. An attacker can gain access by simply knowing a valid UUID and making a request to the unauthenticated endpoint, revealing privileged information. Implementing authentication or removing the endpoint would eliminate the attack surface.