# CVE-2026-121561: Label Studio server-side request forgery in export formats

Source: https://junglewise.ai/threats/cve-2026-121561-label-studio-before-1-23-2-label-studio-sdk-before-2-1-2  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-11-03. Updated: 2026-11-03  
License: CC BY 4.0 with a link back

CVE-2026-121561 · Severity: high · CVSS 7.7 · Published 2026-11-03

Label Studio is an open-source platform for data annotation and labeling used in machine learning workflows. Authenticated users can exploit a server-side request forgery vulnerability to read responses from internal services and capture sensitive credentials like API tokens by manipulating image import and export functionality. This allows attackers with valid accounts to access data they should not be able to reach, including organization owner credentials.

Vendors: [HumanSignal](https://junglewise.ai/threats/vendors/humansignal).

## Quick answers

**Is CVE-2026-121561 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-121561 being exploited in the wild.

**How severe is CVE-2026-121561?** High severity, with a CVSS score of 7.7, published 3 November 2026.

**Which products and versions are affected?** HumanSignal Label Studio before 1.23.2; HumanSignal Label-Studio-Sdk before 2.1.2.

**Is there a fix for CVE-2026-121561?** Yes: fixed in Label Studio 1.23.2, label-studio-sdk 2.1.2. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

**How is CVE-2026-121561 attacked?** It is attacked over the network, in a software library.

**Am I affected by CVE-2026-121561?** Check whether you run HumanSignal Label Studio or HumanSignal Label-Studio-Sdk at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## How to fix

- Fixed in Label Studio 1.23.2, label-studio-sdk 2.1.2

## Technical details

The vulnerability exists in the export format handling logic (YOLO and COCO formats) where user-supplied image URLs are fetched server-side without proper validation or restriction to external sources. Authenticated attackers can craft tasks with data.image pointing to internal service endpoints, then export in image-bearing formats to trigger server-side requests that leak response content. Host header spoofing can be combined with this to intercept credentials or session tokens.

## Affected products

- HumanSignal Label Studio before 1.23.2
- HumanSignal label-studio-sdk before 2.1.2

## Timeline

- 2026-11-03: disclosed

## References

- [https://github.com/HumanSignal/label-studio-sdk/blob/2.1.1/src/label_studio_sdk/_extensions/label_studio_tools/core/utils/io.py](https://github.com/HumanSignal/label-studio-sdk/blob/2.1.1/src/label_studio_sdk/_extensions/label_studio_tools/core/utils/io.py)
- [https://github.com/HumanSignal/label-studio-sdk/blob/2.1.1/src/label_studio_sdk/_extensions/label_studio_tools/core/utils/io.py](https://github.com/HumanSignal/label-studio-sdk/blob/2.1.1/src/label_studio_sdk/_extensions/label_studio_tools/core/utils/io.py)