Server side request forgery protections in GitLab CE/EE versions between 8.4 and 14.4.4, between 14.5.0 and 14.5.2, and between 14.6.0 and 14.6.1 would fail to protect against attacks sending requests to localhost on port 80 or 443 if GitLab was configured to run on a port other than 80 or 443
An issue has been discovered in GitLab CE/EE affecting all versions starting from 8.11 before 14.3.6, all versions starting from 14.4 before 14.4.4, all versions starting from 14.5 before 14.5.2. Under specific condition an unauthorised project member was allowed to delete a protected branches due to a business logic error.
A collision in access memoization logic in all versions of GitLab CE/EE before 14.3.6, all versions starting from 14.4 before 14.4.4, all versions starting from 14.5 before 14.5.2, leads to potential elevated privileges in groups and projects under rare circumstances
Accidental logging of system root password in the migration log in all versions of GitLab CE/EE before 14.2.6, all versions starting from 14.3 before 14.3.4, and all versions starting from 14.4 before 14.4.1 allows an attacker with local file system access to obtain system root-level privileges
In all versions of GitLab CE/EE since version 8.0, an attacker can set the pipeline schedules to be active in a project export so when an unsuspecting owner imports that project, pipelines are active by default on that project. Under specialized conditions, this may lead to information disclosure if the project is imported from an untrusted source.
An information disclosure vulnerability in the GitLab CE/EE API since version 8.9.6 allows a user to see basic information on private groups that a public project has been shared with
In all versions of GitLab CE/EE since version 8.0, access tokens created as part of admin's impersonation of a user are not cleared at the end of impersonation which may lead to unnecessary sensitive info disclosure.
In all versions of GitLab CE/EE since version 7.7, the application may let a malicious user create an OAuth client application with arbitrary scope names which may allow the malicious user to trick unsuspecting users to authorize the malicious client application using the spoofed scope name and description.