This vulnerability stems from incomplete neutralization of untrusted content channeled into GitLab’s AI‑assisted code review prompts, allowing an authenticated user to craft input that can access data from projects for which the user lacks explicit permissions. The result is a breach of confidentiality: the user can read code or issues from unauthorized projects. The weakness is classified as CWE‑1427 and CWE‑74.

Affected systems are GitLab Enterprise Edition. All releases from 19.1 up to and including 19.1.2, and from 19.2 up to and including 19.2.0, are impacted. The issue was fixed in releases 19.1.3, 19.2.1, and later versions.

The CVSS score of 4.3 indicates moderate severity, and the EPSS score of less than 1% suggests a low probability of exploitation at present. The vulnerability is not listed in the CISA KEV catalog. Exploitation requires an authenticated user with access to the AI‑assisted code review feature; the attacker must construct special prompt content to retrieve unauthorised project data. Given the moderate score and low exploit probability, the risk is considered moderate, but organizations should prioritize patching to eliminate the issue.