Reasoning As an Attack Surface: Adaptive Evolutionary CoT Jailbreaks for LLMs

Large Reasoning Models LRMs have demonstrated remarkable capabilities in reasoning and generation tasks and are increasingly deployed in real-world applications. However, their explicit chain-of-thought CoT mechanism introduces new security risks, making them particularly vulnerable to jailbreak...

CVE-2026-5817

The vllm-metal inference backend in Docker Model Runner on macOS unconditionally sets trustremotecode=True when loading model tokenizers, and runs without sandboxing. This causes transformers.AutoTokenizer.frompretrained to import and execute arbitrary Python files included in any model pulled fr...

EUVD-2026-31493

The vllm-metal inference backend in Docker Model Runner on macOS unconditionally sets trustremotecode=True when loading model tokenizers, and runs without sandboxing. This causes transformers.AutoTokenizer.frompretrained to import and execute arbitrary Python files included in any model pulled fr...

SUSE-SU-2026:21756-1 Security update for mcphost

This update for mcphost fixes the following issues - CVE-2025-30153: github.com/getkin/kin-openapi/openapi3filter: Improper Handling of Highly Compressed Data Data Amplification in github.com/getkin/kin-openapi/openapi3filter bsc1264762. - CVE-2025-47913: golang.org/x/crypto/ssh/agent: client...

OPENSUSE-SU-2026:20788-1 Security update for mcphost

This update for mcphost fixes the following issues - CVE-2025-30153: github.com/getkin/kin-openapi/openapi3filter: Improper Handling of Highly Compressed Data Data Amplification in github.com/getkin/kin-openapi/openapi3filter bsc1264762. - CVE-2025-47913: golang.org/x/crypto/ssh/agent: client...

Are Frontier LLMs Ready for Cybersecurity? Evidence for Vertical Foundation Models from Dual-Mode Vulnerability Benchmarks

We evaluate whether frontier LLMs are ready for cybersecurity through a dual-mode benchmark: white-box function-level vulnerability detection VulnLLM-R, across C/Java/Python and black-box web application security testing five production-style applications with 118 ground-truth vulnerabilities...

AI Security Research Should Better Incentivize Defense Research

This work examines an imbalance in artificial intelligence AI security research: the field tends to produce more work on attacking AI systems than on defending them. Drawing on related academic papers, we find biased attack-to-defense ratios across subfields, including federated learning, speech...

An Empirical Evaluation of LLM-Generated Code Security across Prompting Methods

The growing use of Large Language Models LLMs for automated code generation has enhanced software development efficiency, but often at the cost of security. Generated code frequently overlooks critical concerns, leaving it vulnerable to issues such as weak encryption and improper input validation...

Ollama Scanner

This module identifies ollama instances and enumerates the LLM models which have been loaded and are running. Module Options msf use auxiliary/scanner/http/ollamainfo msf auxiliaryollamainfo show actions ...actions... msf auxiliaryollamainfo set ACTION msf auxiliaryollamainfo show options ...show...

EUVD-2026-30420

Cleartext storage of HMAC signing key in Amazon SageMaker Python SDK ModelBuilder/Serve path...

Cleartext storage of HMAC signing key in Amazon SageMaker Python SDK ModelBuilder/Serve path

Summary Amazon SageMaker Python SDK is an open-source library for training and deploying machine learning models on Amazon SageMaker. An issue exists where, under certain circumstances, the ModelBuilder/Serve component stores an HMAC signing key in cleartext as a container environment variable,...

GHSA-M549-QQ94-FVHG LMDeploy: Arbitrary code execution via hardcoded trust_remote_code=True in lmdeploy model initialization

Summary lmdeploy hardcodes trustremotecode=True in multiple HuggingFace model-loading call sites. The affected code paths are in: text lmdeploy/archs.py lmdeploy/utils.py The vulnerable call sites pass trustremotecode=True into HuggingFace Transformers APIs such as AutoConfig.frompretrained,...

PT-2026-42632

Summary lmdeploy hardcodes trust remote code=True in multiple HuggingFace model-loading call sites. The affected code paths are in: text lmdeploy/archs.py lmdeploy/utils.py The vulnerable call sites pass trust remote code=True into HuggingFace Transformers APIs such as AutoConfig.from pretrained,...

Blind Spots in the Guard: How Domain-Camouflaged Injection Attacks Evade Detection in Multi-Agent LLM Systems

Injection detectors deployed to protect LLM agents are calibrated on static, template-based payloads that announce themselves as override directives. We identify a systematic blind spot: when payloads are generated to mimic the domain vocabulary and authority structures of the target document, wh...

Human Vulnerability Assessment in Cybersecurity: A Systematic Literature Review of Methods, Models, and Instruments

In cybersecurity, vulnerability assessment has typically focused on identifying and measuring vulnerabilities within digital assets and technical infrastructures. However, there is growing recognition that this approach alone is inadequate without a structured examination of the human factor, whi...

Security of LLM-Generated Code: A Comparative Analysis

The majority of software developers use or are planning to use Artificial Intelligence AI tools in their development processes. Their top reasons include improving productivity and faster learning. In fact, Large Language Model LLM-generated code is currently in production, including in major tec...

Prompt Overflow: What the Guardrail Inspects Is Not What the Model Infers

Guardrail models a.k.a. safety checkers are widely deployed to screen user inputs before they reach large language models LLMs, serving as a primary defense against prompt injection attacks. Due to strict context constraints, these models handle overlength prompts through truncation or...

Pretraining Data Exposure in Large Language Models: A Survey of Membership Inference, Data Contamination, and Security Implications

Large Language Models LLMs have become the predominant paradigm in NLP, advancing both research and industry. As model sizes and pretraining data grow, concerns about Pretraining Data Exposure PDE increase due to the scale and opacity of training datasets. PDE refers to determining whether specif...

BIT-PYTHON-2026-4224 Stack overflow parsing XML with deeply nested DTD content models

When an Expat parser with a registered ElementDeclHandler parses an inline document type definition containing a deeply nested content model a C stack overflow occurs...

BIT-PYTHON-MIN-2026-4224 Stack overflow parsing XML with deeply nested DTD content models

When an Expat parser with a registered ElementDeclHandler parses an inline document type definition containing a deeply nested content model a C stack overflow occurs...