The vulnerability resides in the CowAgent component of the ChatGPT on WeChat application. An unknown function within the Agent Mode Service fails to enforce authentication, allowing an attacker to execute that function without any credentials. This omission can grant the attacker unauthorized control over the service, potentially exposing sensitive data, altering service behavior, or enabling further exploitation within the host system.

Products affected include the ChatGPT on WeChat CowAgent released by the developer zhayujie. All versions up to 2.0.4 are vulnerable, as documented in the advisory and issue reports. Users of these releases should verify their version and seek an updated release that addresses the authentication bypass.

The CVSS base score of 6.9 indicates moderate to high impact. EPSS data is unavailable, and the vulnerability is not listed in the CISA KEV catalog, suggesting it has not yet become widely exploited. However, because the flaw can be triggered remotely, an attacker who can reach the Agent Mode Service endpoint could potentially exploit it. The lack of authentication makes the attack path simple, but the exact prerequisites (e.g., specific request format) remain undetailed; the available references indicate prevention is currently limited to applying an update.