The flaw resides in the ChatCompletion.create function of freegpt-webui where an authentication check is omitted, enabling an attacker to invoke the endpoint without credentials. The missing check can allow any user with network reachability to interact with the model, potentially exposing sensitive prompts or extracting data via the API. This represents an authentication failure that could result in unauthorized data access or service misuse.

The vulnerability affects the ramon-victor freegpt-webui product. Any deployment using a release up to the commit 098db3dfeb41555c2ca9269df0f13e10ec1c35dc is impacted. Newer releases after this commit are presumed fixed, though the project uses a rolling release model without explicit version tags.

The CVSS score of 6.9 indicates moderate severity. Even though no EPSS score is reported, the vulnerability is publicly disclosed and can be exploited remotely. It is not currently listed in the CISA KEV catalog, but the fact that any authenticated or unauthenticated request can bypass safety checks suggests that attackers could leverage it until a patch is applied.