Taint-Based Code Slicing for LLMs-Based Malicious NPM Package Detection

The increasing sophistication of malware attacks in the npm ecosystem, characterized by obfuscation and complex logic, necessitates advanced detection methods. Recently, researchers have turned their attention from traditional detection approaches to Large Language Models LLMs due to their strong...

Automated Penetration Testing with LLM Agents and Classical Planning

While penetration testing plays a vital role in cybersecurity, achieving fully automated, hands-off-the-keyboard execution remains a significant research challenge. In this paper, we introduce the "Planner-Executor-Perceptor PEP" design paradigm and use it to systematically review existing work a...

LLM-Assisted AHP for Explainable Cyber Range Evaluation

Cyber Ranges CRs have emerged as prominent platforms for cybersecurity training and education, especially for Critical Infrastructure CI sectors that face rising cyber threats. One way to address these threats is through hands-on exercises that bridge IT and OT domains to improve defensive...

LLM-Based Vulnerable Code Augmentation: Generate or Refactor?

Vulnerability code-bases often suffer from severe imbalance, limiting the effectiveness of Deep Learning-based vulnerability classifiers. Data Augmentation could help solve this by mitigating the scarcity of under-represented CWEs. In this context, we investigate LLM-based augmentation for...

nim-pentest-agent

NimPentestAgent Agent autonome de pentest intelligent pour CT...

LLM Causality Analysis Framework

A comprehensive framework for multi-level causality analysis in Large Language Models LLMs, enabling systematic investigation of safety mechanisms and misbehavior detection across token, neuron, layer, and representation levels. Includes the whitepaper 2512.04841.pdf titled SoK: A Comprehensive...

DRUPAL-CONTRIB-2025-119

This modules provides the ability to chat with an AI Agent using a large-language model LLM provider for different purposes. The module doesn’t sufficiently filter LLM responses. This leads to a cross-site scripting XSS vulnerability where an attacker can use prompt injections on user-generated...

AI (Artificial Intelligence) - Moderately critical - Cross-Site Scripting - SA-CONTRIB-2025-119

This modules provides the ability to chat with an AI Agent using a large-language model LLM provider for different purposes. The module doesn’t sufficiently filter LLM responses. This leads to a cross-site scripting XSS vulnerability where an attacker can use prompt injections on user-generated...

CVE-2025-62155

New API is a large language mode LLM gateway and artificial intelligence AI asset management system. Prior to version 0.9.6, a recently patched SSRF vulnerability contains a bypass method that can bypass the existing security fix and still allow SSRF to occur. Because the existing fix only applie...

Malicious code in chat-prompt-logger (PyPI)

--- -= Per source details. Do not edit below this line.=- Source: kam193 f25a736985f5c0bb50156fdc7de61e976b16416f42c44a2682b5ce718401383b The package provides a logger of LLM prompts that at the same time looks for hidden instructions and executes them. --- Category: MALICIOUS - The campaign has...

CVE-2025-33204

NVIDIA NeMo Framework for all platforms contains a vulnerability in the NLP and LLM components, where malicious data created by an attacker could cause code injection. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, information disclosure, and data...

NVIDIA Nemo Framework 代码注入漏洞

NVIDIA Nemo Framework is a framework for building and deploying generative AI models from NVIDIA. A code injection vulnerability exists in NVIDIA Nemo Framework that stems from the presence of malicious data in the NLP and LLM components, which could lead to code injection that could result in co...

Effective Command-Line Interface Fuzzing with Path-Aware Large Language Model Orchestration

Command-line interface CLI fuzzing tests programs by mutating both command-line options and input file contents, thus enabling discovery of vulnerabilities that only manifest under specific option-input combinations. Prior works of CLI fuzzing face the challenges of generating semantics-rich opti...

CVE-2025-62155

The CVE-2025-62155 entry concerns QuantumNous/new-api. A SSRF vulnerability existed prior to version 0.9.6 where the fix only protected the first URL request; an attacker could bypass via a 302 redirect and reach internal/intranet resources. The issue has been addressed in version 0.9.6, accordin...

DUALGUAGE: Automated Joint Security-Functionality Benchmarking for Secure Code Generation

Large language models LLMs and autonomous coding agents are increasingly used to generate software across a wide range of domains. Yet a core requirement remains unmet: ensuring that generated code is secure without compromising its functional correctness. Existing benchmarks and evaluations for...

PT-2025-47975

Name of the Vulnerable Software and Affected Versions New API versions prior to 0.9.6 Description New API is a large language model LLM gateway and artificial intelligence AI asset management system. A Server-Side Request Forgery SSRF condition existed in versions prior to 0.9.6. A previous...

AI teddy bear for kids responds with sexual content and advice about weapons

In testing, FoloToy’s AI teddy bear jumped from friendly chat to sexual topics and unsafe household advice. It shows how easily artificial intelligence can cross serious boundaries. It’s a fair moment to ask whether AI-powered stuffed animals are appropriate for children. It’s easy to get swept u...

CVE-2025-62426

vLLM is an inference and serving engine for large language models LLMs. From version 0.5.5 to before 0.11.1, the /v1/chat/completions and /tokenize endpoints allow a chattemplatekwargs request parameter that is used in the code before it is properly validated against the chat template. With the...

EUVD-2025-198357

vLLM is an inference and serving engine for large language models LLMs. From version 0.5.5 to before 0.11.1, users can crash the vLLM engine serving multimodal models by passing multimodal embedding inputs with correct ndim but incorrect shape e.g. hidden dimension is wrong, regardless of whether...

EUVD-2025-198356

vLLM is an inference and serving engine for large language models LLMs. From version 0.5.5 to before 0.11.1, the /v1/chat/completions and /tokenize endpoints allow a chattemplatekwargs request parameter that is used in the code before it is properly validated against the chat template. With the...