Home

Description

LLaMA-Factory through 0.9.5 contains a remote code execution vulnerability that allows attackers with WebUI access to execute arbitrary Python code by supplying a malicious model path in the Chat or Training interfaces. The application passes user-supplied model path input unvalidated into AutoTokenizer.from_pretrained() and AutoModel.from_pretrained() with a hardcoded trust_remote_code=True parameter, causing the Hugging Face transformers library to fetch and execute arbitrary code from a remote or local model repository with the privileges of the server process.

PUBLISHED Reserved 2026-06-29 | Published 2026-06-30 | Updated 2026-07-14 | Assigner VulnCheck




CRITICAL: 9.3CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N

CRITICAL: 9.8CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H

Problem types

Inclusion of Functionality from Untrusted Control Sphere

Improper Control of Generation of Code ('Code Injection')

Product status

Default status
affected

Any version
affected

Credits

h3nrrrych4u finder

References

gist.github.com/henrrrychau/08d76ec672f42136bbc1449c4f2973f8 (Researcher Disclosure) technical-description exploit

www.vulncheck.com/...ote-code-execution-via-webui-model-path third-party-advisory

cve.org (CVE-2026-58116)

nvd.nist.gov (CVE-2026-58116)

Download JSON