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.

In all versions of GitLab CE/EE since version 11.11, an instance that has the setting to disable Repo by URL import enabled is bypassed by an attacker making a crafted API call.

An issue has been discovered in GitLab affecting all versions starting from 13.8 before 14.0.9, all versions starting from 14.1 before 14.1.4, all versions starting from 14.2 before 14.2.2. Under specialized conditions, an invited group member may continue to have access to a project even after the invited group, which the member was part of, is deleted.

Missing access control in all GitLab versions starting from 13.12 before 14.0.9, all versions starting from 14.1 before 14.1.4, and all versions starting from 14.2 before 14.2.2 with Jira Cloud integration enabled allows Jira users without administrative privileges to add and remove Jira Connect Namespaces via the GitLab.com for Jira Cloud application configuration page

A stored Cross-Site Scripting vulnerability in the Jira integration in all GitLab versions starting from 13.9 before 14.0.9, all versions starting from 14.1 before 14.1.4, and all versions starting from 14.2 before 14.2.2 allows an attacker to execute arbitrary JavaScript code on the victim's behalf via malicious Jira API responses

The project import/export feature in GitLab 8.9 and greater could be used to obtain otherwise private email addresses

An issue has been discovered in GitLab affecting all versions starting from 14.0 before 14.0.9, all versions starting from 14.1 before 14.1.4, all versions starting from 14.2 before 14.2.2. The route for /user.keys is not restricted on instances with public visibility disabled. This allows user enumeration on such instances.

In all versions of GitLab CE/EE since version 8.0, a DNS rebinding vulnerability exists in Fogbugz importer which may be used by attackers to exploit Server Side Request Forgery attacks.

A potential DOS vulnerability was discovered in GitLab starting with version 9.1 that allowed parsing files without authorisation.

In all versions of GitLab EE since version 8.13, an endpoint discloses names of private groups that have access to a project to low privileged users that are part of that project.

In all versions of GitLab CE/EE, provided a user ID, anonymous users can use a few endpoints to retrieve information about any GitLab user.

A stored Reflected Cross-Site Scripting vulnerability in the Jira integration in GitLab version 13.0 up to 14.3.1 allowed an attacker to execute arbitrary javascript code.

In all versions of GitLab CE/EE since version 13.6, it is possible to see pending invitations of any public group or public project by visiting an API endpoint.

In all versions of GitLab CE/EE since version 14.1, an improper access control vulnerability allows users with expired password to still access GitLab through git and API through access tokens acquired before password expiration.

In all versions of GitLab CE/EE since version 8.9, project exports may expose trigger tokens configured on that project.

In all versions of GitLab CE/EE since version 8.15, a DNS rebinding vulnerability in Gitea Importer may be exploited by an attacker to trigger Server Side Request Forgery (SSRF) attacks.

A business logic error in the project deletion process in GitLab 13.6 and later allows persistent access via project access tokens.

A stored Cross-Site Scripting vulnerability in the GitLab Flavored Markdown in GitLab CE/EE version 8.4 and above allowed an attacker to execute arbitrary JavaScript code on the victim's behalf.

Information disclosure from SendEntry in GitLab starting with 10.8 allowed exposure of full URL of artifacts stored in object-storage with a temporary availability via Rails logs.

In all versions of GitLab CE/EE, an attacker with physical access to a user’s machine may brute force the user’s password via the change password function. There is a rate limit in place, but the attack may still be conducted by stealing the session id from the physical compromise of the account and splitting the attack over several IP addresses and passing in the compromised session value from these various locations.