Flowise versions prior to 3.1.0 are vulnerable to an authorization bypass that permits unauthenticated access to sensitive API credentials.

The vulnerability exists in the text-to-speech generation endpoint (POST /api/v1/text-to-speech/generate), which fails to perform proper authorization checks. An unauthenticated attacker can supply a credentialId in the request body to decrypt and access sensitive stored credentials, such as OpenAI or ElevenLabs API keys.

The exposure of API keys poses a significant risk to organizational security, as these credentials can be used to incur unauthorized costs or access private LLM infrastructure. With a CVSS score of 7.5, this high-severity flaw could lead to data exfiltration or service disruption. Unauthorized access to these keys compromises the integrity of the integrated LLM workflows and the associated cloud accounts.

Immediate Action: Update Flowise to version 3.1.0 or later immediately to apply the necessary authorization checks for the text-to-speech endpoint.

Proactive Monitoring: Review API access logs for unusual POST requests to the /api/v1/text-to-speech/generate endpoint, specifically looking for requests that lack associated chatflow identifiers.

Compensating Controls: Deploy a Web Application Firewall (WAF) to restrict access to the /api/v1/text-to-speech/generate endpoint, limiting traffic to trusted internal IP addresses until the patch can be applied.

Public Exploit Available: Unknown

Given the ease of exploitation, which requires no authentication, organizations should prioritize upgrading their Flowise instances to version 3.1.0 as soon as possible. Failure to patch may result in the compromise of sensitive third-party API credentials, leading to financial loss or further unauthorized access. Ensure all development and production environments are covered by this update.