The Drupal AI module has a missing authorization check that allows an attacker to forcefully browse to restricted content. This vulnerability is essentially an authorization bypass exposed through direct URL access or manipulation, enabling unauthorized read access to data controlled by the AI module. The weakness corresponds to CWE‑862, which indicates improper enforcement of access controls.

Affected by the Drupal AI (Artificial Intelligence) module across all installations running any version from 0.0.0 up to and including 1.3.13, as well as versions 1.4.0 through 1.4.8. Any deployment of these specific releases is considered vulnerable until patched.

The CVSS score is 4.6, and the EPSS score is less than 1%, indicating a modest risk profile. The vulnerability is publicly documented on Drupal’s security advisory page and is not listed in the CISA KEV catalog, suggesting it has not yet been exploited in the wild. However, the lack of an authentication bypass means that a remote attacker could exploit it by simply accessing predictable URLs, provided they have network visibility to the Drupal site. The potential impact is unauthorized access to protected content, with the possibility of confidentiality leakage. Because it is an authorization flaw, the risk landscape is primarily limited to environments that deploy the exposed module without additional access controls. The likely attack vector involves an unauthenticated attacker navigating to URLs that rely on the AI module for content delivery. No additional privileges are required beyond network reach, and once an attacker reads restricted content, no further privileges can be escalated with this flaw alone. The absence of post‑exploitation stages indicates that the main threat is data exposure, but the exposed data could serve as a foothold for other attacks.