MEDIUM6.5
PYSEC-2026-2668
MS-Agent vulnerable to Command Injection
Details
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.
Are you affected?
Enter the version of the package you're using.
Affected packages
PyPI/ms-agent
Introduced in:
0No fixed version published yet for ms-agent (pip). Pin to a known-safe version or switch to an alternative.
References
- https://nvd.nist.gov/vuln/detail/CVE-2026-2256[ADVISORY]
- https://github.com/Itamar-Yochpaz/CVE-2026-2256-PoC[WEB]
- https://github.com/modelscope/ms-agent[PACKAGE]
- https://medium.com/@itamar.yochpaz/cve-2026-2256-from-ai-prompt-to-full-system-compromise-a4114c718326[WEB]
- https://www.hiddenlayer.com/research/indirect-prompt-injection-of-claude-computer-use[WEB]
- https://www.kb.cert.org/vuls/id/431821[WEB]
- https://pypi.org/project/ms-agent[PACKAGE]
- https://github.com/advisories/GHSA-4gc2-344q-r2rw[ADVISORY]