CVAT is an open source interactive video and image annotation tool for computer vision. From 2.68.0 until 2.70.0, the audio-task annotation guide renderer in cvat-ui/src/audio/components/annotation-page/audio-workspace/top-bar/audio-right-group.tsx passes attacker-controlled guide Markdown to MDEditor without the rehype-sanitize plugin. A user who can create or edit an annotation guide can store malicious JavaScript that executes when another user opens the guide. The script can issue arbitrary CVAT requests with the victim user's privileges. This issue is fixed in version 2.70.0.
CVAT is an open source interactive video and image annotation tool for computer vision. From 2.17.0 until 2.72.0, a user with the Worker role can use predictable task-based request IDs with the lambda request retrieve and destroy endpoints to view automatic annotation requests for tasks or jobs the user cannot access and cancel requests initiated by other users. This issue is fixed in version 2.72.0.
CVAT is an open source interactive video and image annotation tool for computer vision. From 2.17.0 until 2.72.0, a user with write access to a CVAT job can submit a batch automatic annotation request to RequestViewSet.create with inconsistent task and job IDs, and because the task ID determines the single active request slot, block automatic annotation for another task whose ID is known. This issue is fixed in version 2.72.0.
CVAT versions 2.5.0 through 2.66.0 contain a cross-site scripting (XSS) vulnerability related to how annotation guide assets are served. An attacker can influence the Content-Type header of files attached to annotation guides, causing a victim's browser to interpret malicious files as HTML and execute embedded JavaScript. This vulnerability has been fixed in version 2.67.0.
CVE-2026-47682 is a path traversal vulnerability in the CVAT open source annotation tool. Versions from 1.6.0 up to but not including 2.65.0 allow an attacker with write access to a configured cloud storage, or the ability to add new cloud storages, to overwrite arbitrary files on the server filesystem. This vulnerability has been fixed in version 2.65.0.
CVAT is an open source interactive video and image annotation tool for computer vision. From 2.5.0 to 2.63.0, an attacker who is able to create or edit an annotation guide on a task is able to add malicious JavaScript code, which will then run in the browser of anyone who opens this annotation guide. This code will be able to make arbitrary requests to CVAT with the victim user's privileges. This vulnerability is fixed in 2.64.0.
CVAT is an open source interactive video and image annotation tool for computer vision. In versions 1.0.0 through 2.54.0, users that have the staff status may freely change their permissions, including giving themselves superuser status and joining the admin group, which gives them full access to the data in the CVAT instance. Version 2.55.0 fixes the issue. As a workaround, review the list of users with staff status and revoke it from any users that are not expected to have superuser privileges.
CVAT is an open source interactive video and image annotation tool for computer vision. In versions 2.2.0 through 2.54.0, an attacker is able to execute arbitrary JavaScript in a victim user's CVAT UI session, provided that they are able to create a maliciously crafted label in a CVAT task or project, then get the victim user to either edit that label, or view a shape that refers to that label; and/or get the victim user to upload a maliciously crafted SVG image when configuring a skeleton. This gives the attacker temporary access to all CVAT resources that the victim user can access. Version 2.55.0 fixes the issue.
CVAT is an open source interactive video and image annotation tool for computer vision. In versions 2.8.1 through 2.52.0, an attacker with an account on a CVAT instance is able to retrieve the contents of any file system directory accessible to the CVAT server. The exposed information is names of contained files and subdirectories. The contents of files are not accessible. Version 2.53.0 contains a patch. No known workarounds are available.
CVAT is an open source interactive video and image annotation tool for computer vision. In versions 2.4.0 through 2.48.1, a malicious CVAT user with at least the User global role may create files in the root of the mounted file share, or overwrite existing files. If no file share is mounted, the user will be able to create files in the share directory of the import worker container, potentially filling up disk space. This issue is fixed in version 2.49.0.
CVAT is an open source interactive video and image annotation tool for computer vision. Versions 2.2.0 through 2.39.0 have no validation during the import process of a project or task backup to check that the filename specified in the query parameter refers to a TUS-uploaded file belonging to the same user. As a result, if an attacker with a CVAT account and a `user` role knows the filenames of other users' uploads, they could potentially access and steal data by creating projects or tasks using those files. This issue does not affect annotation or dataset TUS uploads, since in this case object-specific temporary directories are used. Users should upgrade to CVAT 2.40.0 or a later version to receive a patch. No known workarounds are available.
Computer Vision Annotation Tool (CVAT) is an interactive video and image annotation tool for computer vision. In versions starting from 2.4.0 to before 2.38.0, an authenticated CVAT user may be able to retrieve the IDs and names of all tasks, projects, labels, and the IDs of all jobs and quality reports on the CVAT instance. In addition, if the instance contains many resources of a particular type, retrieving this information may tie up system resources, denying access to legitimate users. This issue has been patched in version 2.38.0.