TorchServe Potentially Exposed to Remote Code Execution
Israeli-based security company Oligo has uncovered multiple vulnerabilities in TorchServe, the tool used to serve PyTorch models, that could allow an attacker to run arbitrary code on vulnerable systems. The vulnerabilities have been promptly fixed in TorchServe version 0.82.
Zoom Open-sources New Vulnerability Impact Scoring System VISS
Zoom Vulnerability Impact Scoring System, or VISS for short, aims to help organizations enforce security measures based on a new approach to vulnerability scoring that prioritizes actual demonstrated impact over theoretical security impact possibilities.
GitHub CodeQL Code Scanning Now Supports Setting a Threat Model
GitHub has recently extended its CodeQL-based code scanner by adding the possibility to specify the desired threat model. The new feature is available in beta for the Java language.
LeftoverLocals May Leak LLM Responses on Apple, Qualcomm, and AMD GPUs
Security firm Trail of Bits disclosed a vulnerability allowing malicious actors to recover data from GPU local memory on Apple, Qualcomm, AMD, and Imagination GPUs. Dubbed LeftoverLocals, the vulnerability affects any application using the GPU, including Large Language Models (LLMs) and machine learning (ML) models.