Skip to content
Submit incident
AII-1002-46543
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.

CVAT broken access control in PATCH endpoints (CVE-2024-47172)

CVAT, an open-source computer vision annotation tool, contained a broken access control vulnerability in PATCH endpoints that allowed authenticated attackers to view and modify project, task, job and membership resources. The issue was patched in version 2.19.1.

30 September 2024 Β· CVAT AI Β· Security and cyberattacks
30 September 2024
AII-1002-46543View source β†—
Quick facts
Organization
CVAT AI (subject)
Category
Security and cyberattacks
Source
github.com
Date
30 September 2024
Drafted by
Raidu Incident Detection

What happened

CVAT versions 2.0.0 through 2.19.0 contained broken access control in several PATCH endpoints affecting projects, tasks, jobs and memberships resources.

An attacker with a CVAT account could retrieve information about any project, task, job or membership resource on the CVAT instance by making PATCH requests, retrieving the same information as GET requests would return.

The attacker could also alter the default source and target storage associated with any project or task.

Reported impact

The vulnerability allowed unauthorized access to information and modification of storage settings for any resource on the CVAT instance.

Index analysis

The vulnerability was a broken access control issue (CWE-863) in authorization checks for PATCH requests, with CVSS 3.1 score of 5.4 (Moderate severity).

The attack vector was network-based, requiring low privileges and user interaction, with low impact on confidentiality and integrity.

Governance controls

CVAT patched the vulnerability in version 2.19.1 by implementing proper authorization checks for PATCH requests, including requiring view permissions and adding a new scope for storage modifications.

Classification

Organization
CVAT AI (subject)
Category
Security and cyberattacks
Industry
IT infrastructure and hosting
Harm status
Potential harm
Intent
Unintentional
Timing
Post-deployment
Failure modes
Data leak
AI systems
CVAT

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

github.com
https://github.com/cvat-ai/cvat/security/advisories/GHSA-gxhm-hg65-5gh2