A high-severity authentication bypass vulnerability in Flowise allows unauthenticated attackers to modify or delete sensitive data within OpenAI Assistants Vector Stores.

All CRUD endpoints related to the OpenAI Assistants Vector Store lack necessary authentication middleware. This allows an unauthenticated remote attacker to perform unauthorized operations, including the modification or deletion of data managed within the Flowise interface.

With a CVSS score of 8.8, this vulnerability poses a severe threat to the integrity and availability of AI-driven workflows. An attacker could destroy valuable training data or inject malicious information into vector stores, leading to corrupted model behavior and potential loss of intellectual property.

Immediate Action: Upgrade Flowise to version 3.1.2 or later to implement the required authentication middleware.

Proactive Monitoring: Review access logs for the Flowise administrative interface and monitor for unauthorized API requests targeting the Vector Store endpoints.

Compensating Controls: Implement network-level access controls to restrict access to the Flowise management interface to authorized internal IP addresses only.

Public Exploit Available: false

The absence of authentication on critical API endpoints makes this an urgent security priority. All users of Flowise must upgrade to version 3.1.2 immediately to secure the platform against unauthorized data manipulation. Restricting network access to the application is a recommended secondary measure until the update is fully deployed.