Lucene search
K

6 matches found

Github Security Blog
Github Security Blog
added last week50 views

vLLM: Security Check Bypass via assert Statement in Activation Function Loading Allows Arbitrary Code Execution

Summary An assert-based security check in vLLM's activation function loading allows any unauthenticated attacker to achieve arbitrary code execution on the server by publishing a malicious HuggingFace model, when vLLM runs in Python optimized mode python -O or PYTHONOPTIMIZE=1. Details vLLM uses ...

8.1CVSS7.7AI score0.02469EPSS
Exploits0References4Affected Software1
Positive Technologies
Positive Technologies
added 2026/06/16 12:0 a.m.9 views

PT-2026-50140

Name of the Vulnerable Software and Affected Versions vLLM versions prior to 0.22.0 Description An assert-based security check in the activation function loading process allows an unauthenticated attacker to achieve arbitrary code execution on the server. This occurs when vLLM is run in Python...

7.5CVSS7.4AI score0.00054EPSS
Exploits0References6
RedhatCVE
RedhatCVE
added 2026/06/11 2:59 a.m.7 views

CVE-2026-46432

LMDeploy is a toolkit for compressing, deploying, and serving large language models. In versions 0.12.3 and prior, LMDeploy is vulnerable to arbitrary code execution through hardcoded "trustremotecode=True" in multiple HuggingFace model-loading call sites. At time of publication, there are no...

7.8CVSS6.2AI score0.00142EPSS
Exploits0References1
EUVD
EUVD
added 2026/06/09 11:5 p.m.7 views

EUVD-2026-35873

LMDeploy is a toolkit for compressing, deploying, and serving large language models. In versions 0.12.3 and prior, LMDeploy is vulnerable to arbitrary code execution through hardcoded "trustremotecode=True" in multiple HuggingFace model-loading call sites. At time of publication, there are no...

7.8CVSS6.2AI score0.00142EPSS
Exploits0References1
OSV
OSV
added 2026/04/22 3:31 p.m.3 views

GHSA-RXPQ-XGQX-FR7P InstructLab Includes Functionality from Untrusted Control Sphere

A flaw was found in InstructLab. The linuxtrain.py script hardcodes trustremotecode=True when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary Python code execution by convincing a user to run ilab train/download/generate with a specially crafted malicious model...

8.8CVSS6.2AI score0.00353EPSS
Exploits0References4
Positive Technologies
Positive Technologies
added 2026/04/22 12:0 a.m.4 views

PT-2026-34336

A flaw was found in InstructLab. The linux train.py script hardcodes trust remote code=True when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary Python code execution by convincing a user to run ilab train/download/generate with a specially crafted malicious...

8.8CVSS6.3AI score0.00353EPSS
Exploits0References3
Rows per page
Query Builder