Auth bypass in Cvat Computer Vision Annotation Tool
CVE-2025-54573
CVAT is an open source interactive video and image annotation tool for computer vision. In versions 1.1.0 through 2.41.0, email verification was not enforced when using Basic HTTP Authentication. As a result, users could create accounts using fake email addresses and use the product as verified users. Additionally, the missing email verification check leaves the system open to bot signups and further usage. CVAT 2.42.0 and later versions contain a fix for the issue. CVAT Enterprise customers have a workaround available; those customers may disable registration to prevent this issue.
Vulnerability class: Broken Authentication
Published · last modified .
CVSS v3 metric
CVSS v3 base score 4.3 (Medium). Vector: CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L.
EPSS exploit prediction
EPSS 0.003, 23.3th percentile (scored ); unchanged over 30 days. About EPSS.
Affected products
- Cvat Computer Vision Annotation Tool
- Cvat-Ai Cvat — versions >= 1.1.0, < 2.42.0