FlashAttention 安全漏洞

FlashAttention is an efficient and memory-efficient attention mechanism implementation tool open-sourced by Dao AI Lab. There is a security vulnerability in FlashAttention; this vulnerability stems from the training script registering the Python eval function as a Hydra configuration parser, whic...

PT-2026-39638

Name of the Vulnerable Software and Affected Versions flash-attention training framework versions prior to commit e724e2588cbe754beb97cf7c011b5e7e34119e62 Description An insecure deserialization issue exists in the checkpoint loading mechanism. The load checkpoint function in checkpoint.py and th...

CVE-2026-31253

The flash-attention training framework thru commit e724e2588cbe754beb97cf7c011b5e7e34119e62 2025-13-04 contains an insecure deserialization vulnerability CWE-502 in its checkpoint loading mechanism. The loadcheckpoint function in checkpoint.py and the checkpoint loading code in eval.py use...

CVE-2026-31254

The CVE-2026-31254 entry concerns the flash-attention project commit e724e2588cbe754beb97cf7c011b5e7e34119e62 (2025-04-13). A code-injection flaw (CWE-94) exists in the training script where Python’s eval() is registered as a Hydra config resolver under the name eval, enabling arbitrary code exec...

Attention Is Where You Attack

Safety-aligned large language models rely on RLHF and instruction tuning to refuse harmful requests, yet the internal mechanisms implementing safety behavior remain poorly understood. We introduce the Attention Redistribution Attack ARA, a white-box adversarial attack that identifies...

Vulnerability Identification by Harnessing Inter-Connected Multi-Source Information

The utilization of third-party open-source libraries is widespread in modern software development. Due to the dependency relationships, vulnerabilities within open-source libraries pose significant security threats to downstream software. However, the library vulnerabilities are usually implicitl...

DeepGuard Secure Code Generation

Large Language Models LLMs for code generation can replicate insecure patterns from their training data. To mitigate this, a common strategy for security hardening is to fine-tune models using supervision derived from the final transformer layer. However, this design may suffer from a final-layer...

Jailbreaking the Matrix: Nullspace Steering for Controlled Model Subversion

Large language models remain vulnerable to jailbreak attacks -- inputs designed to bypass safety mechanisms and elicit harmful responses -- despite advances in alignment and instruction tuning. We propose Head-Masked Nullspace Steering HMNS, a circuit-level intervention that i identifies attentio...

Hybrid ResNet-1D-BiGRU with Multi-Head Attention for Cyberattack Detection in Industrial IoT Environments

This study introduces a hybrid deep learning model for intrusion detection in Industrial IoT IIoT systems, combining ResNet-1D, BiGRU, and Multi-Head Attention MHA for effective spatial-temporal feature extraction and attention-based feature weighting. To address class imbalance, SMOTE was applie...

Towards Predicting Multi-Vulnerability Attack Chains in Software Supply Chains from Software Bill of Materials Graphs

Software supply chain security compromises often stem from cascaded interactions of vulnerabilities, for example, between multiple vulnerable components. Yet, Software Bill of Materials SBOM-based pipelines for security analysis typically treat scanner findings as independent per-CVE Common...

GMA-SAWGAN-GP: A Novel Data Generative Framework to Enhance IDS Detection Performance

Intrusion Detection System IDS is often calibrated to known attacks and generalizes poorly to unknown threats. This paper proposes GMA-SAWGAN-GP, a novel generative augmentation framework built on a Self-Attention-enhanced Wasserstein GAN with Gradient Penalty WGAN-GP. The generator employs...

A Novel Solution for Zero-Day Attack Detection in IDS Using Self-Attention and Jensen-Shannon Divergence in WGAN-GP

The increasing sophistication of cyber threats, especially zero-day attacks, poses a significant challenge to cybersecurity. Zero-day attacks exploit unknown vulnerabilities, making them difficult to detect and defend against. Existing approaches patch flaws and deploy an Intrusion Detection Syst...

How the Graph Construction Technique Shapes Performance in IoT Botnet Detection

The increasing incidence of IoT-based botnet attacks has driven interest in advanced learning models for detection. Recent efforts have focused on leveraging attention mechanisms to model long-range feature dependencies and Graph Neural Networks GNNs to capture relationships between data instance...

A Lightweight Defense Mechanism against Next Generation of Phishing Emails Using Distilled Attention-Augmented BiLSTM

The current generation of large language models produces sophisticated social-engineering content that bypasses standard text screening systems in business communication platforms. Our proposed solution for mail gateway and endpoint deception detection operates in a privacy-protective manner whil...

The Trigger in the Haystack: Extracting and Reconstructing LLM Backdoor Triggers

Detecting whether a model has been poisoned is a longstanding problem in AI security. In this work, we present a practical scanner for identifying sleeper agent-style backdoors in causal language models. Our approach relies on two key findings: first, sleeper agents tend to memorize poisoning dat...

Cascaded Vulnerability Attacks in Software Supply Chains

Most of the current software security analysis tools assess vulnerabilities in isolation. However, sophisticated software supply chain security threats often stem from cascaded vulnerability and security weakness chains that span dependent components. Moreover, although the adoption of Software...

FOCA: Multimodal Malware Classification Via Hyperbolic Cross-Attention

In this work, we introduce FOCA, a novel multimodal framework for malware classification that jointly leverages audio and visual modalities. Unlike conventional Euclidean-based fusion methods, FOCA is the first to exploit the intrinsic hierarchical relationships between audio and visual...

Operational Runtime Behavior Mining for Open-Source Supply Chain Security

Open-source software OSS is a critical component of modern software systems, yet supply chain security remains challenging in practice due to unavailable or obfuscated source code. Consequently, security teams often rely on runtime observations collected from sandboxed executions to investigate...

SUSE CVE-2023-54299

In the Linux kernel, the following vulnerability has been resolved: usb: typec: bus: verify partner exists in typecaltmodeattention Some usb hubs will negotiate DisplayPort Alt mode with the device but will then negotiate a data role swap after entering the alt mode. The data role swap causes the...

EUVD-2023-60505

In the Linux kernel, the following vulnerability has been resolved: usb: typec: bus: verify partner exists in typecaltmodeattention Some usb hubs will negotiate DisplayPort Alt mode with the device but will then negotiate a data role swap after entering the alt mode. The data role swap causes the...