CVE-2026-2256

MEDIUM

6.5

Description

A command injection vulnerability in ModelScope's ms-agent versions v1.6.0rc1 and earlier exists, allowing an attacker to execute arbitrary operating system commands through crafted prompt-derived input.

CVE Details

CVSS v3.1 Score6.5

SeverityMEDIUM

CVSS VectorCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:N

Attack VectorNETWORK

ComplexityLOW

Privileges RequiredNONE

User InteractionNONE

Published3/2/2026

Last Modified3/3/2026

Sourcenvd

Honeypot Sightings0

Weaknesses (CWE)

CWE-77

References

https://github.com/Itamar-Yochpaz/CVE-2026-2256-PoC(cret@cert.org)

https://github.com/modelscope/ms-agent(cret@cert.org)

https://medium.com/@itamar.yochpaz/cve-2026-2256-from-ai-prompt-to-full-system-compromise-a4114c718326(cret@cert.org)

https://www.hiddenlayer.com/research/indirect-prompt-injection-of-claude-computer-use(cret@cert.org)

https://www.kb.cert.org/vuls/id/431821(af854a3a-2127-422b-91ae-364da2661108)

IOC Correlations

No correlations recorded

This product uses data from the NVD API but is not endorsed or certified by the NVD.