CVE-2026-7482

CVE-2026-7482: Ollama GGUF Heap OOB Read Reproduction This repository contains my local reproduction script for CVE-2026-7482, a heap out-of-bounds read in vulnerable Ollama GGUF loading and quantization paths. The important result from this work is narrow: I was able to make the Heap OOB conditi...

CVE-2026-1801C

CVE-2026-1801C QUANTUM-SHIFT / CVE-2026-180A7 BAL-JUMP: Static Analysis of Movement Input Heuristics in Source 2 server.dll Abstract This paper presents a static reverse-engineering analysis of two client-side movement verification routines implemented in the Counter-Strike 2 engine server.dll: t...

Improper Validation of Specified Quantity in Input

Overview vllm is an A high-throughput and memory-efficient inference and serving engine for LLMs Affected versions of this package are vulnerable to Improper Validation of Specified Quantity in Input in allreducermsfusion.py when registering pattern replacements for mixed-dtype allreduce RMSNorm...

Towards Understanding LLM-Based Log Anomaly Detection: An Empirical Study of Performance, Efficiency, and Robustness

Large language models LLMs have demonstrated promising performance in log anomaly detection, yet how their adaptation strategies, architectures, and deployment configurations affect detection effectiveness remains insufficiently understood. To investigate these factors, we conduct a systematic...

Reliable Federated TinyML Deployment for IoT Security

The growing deployment of Internet of Things IoT devices has increased the need for privacy-preserving intrusion detection systems that operate directly on resource-constrained hardware. Federated Learning enables collaborative model training without sharing raw data, but conventional federated...

PT-2026-95545

Name of the Vulnerable Software and Affected Versions LMDeploy versions 0.12.1 through 0.12.2 Description Code injection is possible when loading a malicious HuggingFace model. The issue occurs because the quant dtype value from the model's quantization config is passed to the eval function in...

AGENTQ: Quantization-Conditioned Backdoor Attacks on LLM Agents

Quantization is one of the default deployment paths for open-weight LLM agents, but it is not behavior-preserving: an adversary can release a full-precision checkpoint that passes audits yet misbehaves once quantized, termed as quantization-conditioned attack QCA. Prior QCA work targets free-text...

Variational Probabilistic Quantization for Secret Key Generation

Secret key generation from correlated observations at Alice and Bob, in the presence of an eavesdropper Eve, underpins physical-layer security. Classical pipelines quantize by hand, amplify privacy afterwards, and optimize no objective tied to a key rate. We propose Variational Probabilistic...

Bit-Flip Attacks on Vision-Language-Action Models: Action-Decoding Architecture Shapes the Vulnerability

Quantized Vision-Language-Action VLA models expose a weight-fault surface: Rowhammer-style faults can corrupt deployed INT8 bits. We present the first bit-flip attack on a VLA: a few gradient-selected flips reduce closed-loop success to $0%$, while hundreds of random flips are harmless. Across...

Taming the Security-Energy Paradox: A Green AI Approach to Optimized Android Malware Detection

An increase in advanced Android malware requires the use of deep learning models, which can run on Android devices. But there is a trade-off between security and energy use, as strong detection models can drain the battery of devices fast. This work tests different Multi-Layer Perceptron MLP mode...

PYSEC-2026-3199 TensorFlow vulnerable to `CHECK` fail in `FakeQuantWithMinMaxVarsPerChannelGradient`

Impact When tf.quantization.fakequantwithminmaxvarsperchannelgradient receives input min or max of rank other than 1, it gives a CHECK fail that can trigger a denial of service attack. python import tensorflow as tf arg0=tf.random.uniformshape=1,1, dtype=tf.float32, maxval=None...

TensorFlow vulnerable to `CHECK` fail in `FakeQuantWithMinMaxVarsPerChannelGradient`

ImpactWhen tf.quantization.fakequantwithminmaxvarsperchannelgradient receives input min or max of rank other than 1, it gives a CHECK fail that can trigger a denial of service attack.pythonimport tensorflow as tfarg0=tf.random.uniformshape=1,1, dtype=tf.float32,...

TensorFlow vulnerable to `CHECK` fail in `FakeQuantWithMinMaxVarsGradient`

ImpactWhen tf.quantization.fakequantwithminmaxvarsgradient receives input min or max that is nonscalar, it gives a CHECK fail that can trigger a denial of service attack.pythonimport tensorflow as tfimport numpy as np arg0=tf.constantvalue=np.random.randomsize=2, 2, shape=2, 2,...

PYSEC-2026-3234 TensorFlow vulnerable to `CHECK` fail in `FakeQuantWithMinMaxVarsGradient`

Impact When tf.quantization.fakequantwithminmaxvarsgradient receives input min or max that is nonscalar, it gives a CHECK fail that can trigger a denial of service attack. python import tensorflow as tf import numpy as np arg0=tf.constantvalue=np.random.randomsize=2, 2, shape=2, 2, dtype=tf.float...

PYSEC-2026-3306 TensorFlow vulnerable to `CHECK` fail in `FakeQuantWithMinMaxVarsPerChannel`

Impact If FakeQuantWithMinMaxVarsPerChannel is given min or max tensors of a rank other than one, it results in a CHECK fail that can be used to trigger a denial of service attack. python import tensorflow as tf numbits = 8 narrowrange = False inputs = tf.constant0, shape=4, dtype=tf.float32 min ...

TensorFlow segfault TFLite converter on per-channel quantized transposed convolutions

ImpactWhen converting transposed convolutions using per-channel weight quantization the converter segfaults and crashes the Python process.pythonimport tensorflow as tfclass QuantConv2DTransposedtf.keras.layers.Layer: def buildself, inputshape: self.kernel = self.addweight"kernel", 3, 3,...

PYSEC-2026-3291 TensorFlow segfault TFLite converter on per-channel quantized transposed convolutions

Impact When converting transposed convolutions using per-channel weight quantization the converter segfaults and crashes the Python process. python import tensorflow as tf class QuantConv2DTransposedtf.keras.layers.Layer: def buildself, inputshape: self.kernel = self.addweight"kernel", 3, 3,...

TensorFlow segfault TFLite converter on per-channel quantized transposed convolutions

ImpactWhen converting transposed convolutions using per-channel weight quantization the converter segfaults and crashes the Python process.pythonimport tensorflow as tfclass QuantConv2DTransposedtf.keras.layers.Layer: def buildself, inputshape: self.kernel = self.addweight"kernel", 3, 3,...

Securing LLMs in the Wild: Privacy and Security Challenges at the Edge

Large Language Models LLMs are rapidly moving from research settings into the wild, deployed on enterprise infrastructure, personal devices, and edge platforms. While cloud deployments offer scalable compute, concerns over data sovereignty, compliance, latency, and third-party dependence are...

PT-2026-59905

Impact When converting transposed convolutions using per-channel weight quantization the converter segfaults and crashes the Python process. python import tensorflow as tf class QuantConv2DTransposedtf.keras.layers.Layer: def buildself, input shape: self.kernel = self.add weight"kernel", 3, 3,...