Words Speak Louder Than Code: Investigating Cognitive Heuristics in LLM-Based Code Vulnerability Detection

Researchers and practitioners increasingly apply Large Language Models LLMs for automated vulnerability detection. Recent work has shown that LLMs are susceptible to the same cognitive heuristics that bias human judgment. Yet, no work has investigated whether these heuristics affect a model's...

Cybersecurity Is the True Frontier for Generative AI Success or Failure

Cybersecurity is a real-life test-bed for many machine learning problems at once, especially when considering modern strides in using Large Language Models LLMs to automate processes as "agents.'' Cybersecurity workflows require orchestrating hundreds of standard and bespoke tools through various...

LLM Agents Security Duality: A Comprehensive Survey of Self-Security and Empowered Cybersecurity

Large language model LLM agents are rapidly being integrated into real-world systems. Their autonomy and tool-use capabilities generate substantial value while simultaneously expanding the security attack surface. This survey provides a comprehensive overview of the opportunities and challenges o...

Beyond IOCs: AI-enabled threat intelligence

Welcome to this week's Threat Source newsletter. The issue of AI in cybersecurity is often portrayed as a binary choice: either a force multiplier for our adversaries, or a tool bringing professional obsolescence. The reality is more nuanced. While AI certainly brings some advantage to attackers,...

CrypFormBench: Benchmarking Formal Analysis Capability of Large Language Models for Cryptographic Schemes

Manual formal analysis of cryptographic schemes is labor-intensive and requires substantial expertise. While model-checking tools e.g., Scyther and Tamarin and computational-security tools e.g., CryptoVerif and EasyCrypt improve the automation of security proofs, they still rely on experts to...

HelpBench: Assessing the Ability of LLMs to Provide Privacy, Safety, and Security Advice

This paper introduces HelpBench, a benchmark for assessing whether LLMs are capable of providing accurate help in response to questions about digital privacy, safety, and security. We curated 450 questions representing authentic user situations and developed rubrics for each question to evaluate...

PYSEC-2026-2300

vLLM is an inference and serving engine for large language models LLMs. Prior to 0.22.0, 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 vLL...

EUVD-2026-38400

vLLM is an inference and serving engine for large language models LLMs. From 0.5.5 until 0.23.1rc0, integer truncation of tensor dimensions in vLLM's GGUF dequantize kernels csrc/quantization/gguf/ggufkernel.cu causes partial tensor processing. The output tensor is allocated at full size via...

CVE-2026-53923

Summary of CVE-2026-53923 : The vulnerability affects vLLM (GGUF dequantize kernels) where integer truncation of tensor dimensions causes partially filled output tensors. From 0.5.5 up to 0.23.1rc0, the code allocates the full output tensor (torch::empty) but the CUDA kernel processes only a trun...

PT-2026-51418

Name of the Vulnerable Software and Affected Versions vLLM versions prior to 0.22.1 Description vLLM is an inference and serving engine for large language models. The Dockerfile is susceptible to a dependency confusion attack involving the flashinfer-jit-cache package. This occurs because the...

CNnotator: LLM-Guided Memory Safety Annotation Synthesis

Memory safety errors account for a large proportion of security bugs in systems written in C; modern languages such as Java and Rust prevent such bugs because they are memory-safe by design. To migrate systems to safer languages or identify memory errors, we must first determine how legacy code...

Stop Treating Your LLMs Like Web Servers

...

Multi-View Decompilation for LLM-Based Malware Classification

Malware analysts often inspect compiled binaries through decompiled pseudo-C, when source code is unavailable. Recent work suggests that large language models LLMs can assist this process by classifying decompiled code as benign or malicious, but existing pipelines typically rely on a single...

Microsegmentation: Your Digital First Responder to LLM Threats

...

AttackonCTF: Defending Hardware Security Competition Benchmarks in the Age of LLMs

Hardware security competitions such as HackTheSilicon serve as benchmarking platforms for evaluating vulnerability detection methods and for training humans and AI. However, our study reveals that LLMs threaten their validity. Instead of genuine security reasoning, detectors exploit a diff-style...

Data-Centric Benchmarking of Exploit Generation in LLMs: Understanding the Impact of Fine-Tuning

We study the task of CVE-conditioned exploit generation, where a model drafts proof-of-concept PoC exploits given software vulnerability context. We adopt a data-centric approach, constructing a high-quality dataset via multi-stage preprocessing and introducing a scalable evaluation framework wit...

Ollama System Data Compromise Check Missing

The vulnerability of the Ollama system for running and managing large language models is related to the loading of code without checking its integrity. Exploiting this vulnerability allows a remote attacker to execute arbitrary code...

PI-Hunter: Automated Red-Teaming for Exposing and Localizing Prompt Injections

Large Language Models LLMs are rapidly evolving into agentic systems that interact with external tools and environments, introducing new security risks such as indirect prompt injection attacks through untrusted external sources. Existing defenses mainly focus on blocking malicious content at...

Only 10% of SOCs Say They’re Getting Excellent Value From AI. Here’s What the Second Wave Has to Deliver

Eighteen months ago, the AI SOC was a marketing line. Today it's a budget item. The category has crossed over from interesting to inevitable, with billions of dollars now flowing into AI-powered security operations platforms, agentic SOC tools, and AI co-pilots built into every layer of the...

Steering LLM Viewpoints through Fabricated Evidence Injection

As chatbots increasingly influence daily decision-making, their potential to produce misleading responses poses substantial risks to users. This paper investigates a critical cognitive vulnerability in LLMs: their tendency to uncritically trust external context when presented with fabricated...