Security Concerns for Large Language Models: a Survey

Large Language Models LLMs such as GPT-4 and its recent iterations, Google's Gemini, Anthropic's Claude 3 models, and xAI's Grok have caused a revolution in natural language processing, but their capabilities also introduce new security vulnerabilities. In this survey, we provide a comprehensive...

SafeGenes: Evaluating the Adversarial Robustness of Genomic Foundation Models

Genomic Foundation Models GFMs, such as Evolutionary Scale Modeling ESM, have demonstrated significant success in variant effect prediction. However, their adversarial robustness remains largely unexplored. To address this gap, we propose SafeGenes: a framework for Secure analysis of genomic...

Adversarial Threat Vectors and Risk Mitigation for Retrieval-Augmented Generation Systems

Retrieval-Augmented Generation RAG systems, which integrate Large Language Models LLMs with external knowledge sources, are vulnerable to a range of adversarial attack vectors. This paper examines the importance of RAG systems through recent industry adoption trends and identifies the prominent...

CVE-2025-5276

Versions of the package mcp-markdownify-server before 1.0.0 are vulnerable to Server-Side Request Forgery SSRF via the Markdownify.get function. An attacker can craft a prompt that, once accessed by the MCP host, can invoke the webpage-to-markdown, bing-search-to-markdown, and youtube-to-markdown...

Hijacking Large Language Models Via Adversarial In-Context Learning

In-context learning ICL has emerged as a powerful paradigm leveraging LLMs for specific downstream tasks by utilizing labeled examples as demonstrations demos in the preconditioned prompts. Despite its promising performance, crafted adversarial attacks pose a notable threat to the robustness of...

PT-2025-23167 · Blackmagic Design · Davinci Resolve

Name of the Vulnerable Software and Affected Versions: DaVinci Resolve versions prior to the fixed version Description: The issue is related to the use of entitlement "com.apple.security.cs.disable-library-validation" and the lack of launch and library load constraints, allowing a local attacker...

PT-2025-23141

Name of the Vulnerable Software and Affected Versions: mcp-markdownify-server versions all Description: The issue allows an attacker to craft a prompt that, once accessed by the MCP host, will enable it to read arbitrary files from the host running the server via the get-markdown-file tool...

Important: thunderbird

Issue Overview: Through a series of popup and window.print calls, an attacker can cause a window to go fullscreen without the user seeing the notification prompt, resulting in potential user confusion or spoofing attacks. This vulnerability affects Firefox ESR 102.5, Thunderbird 102.5, and Firefo...

GHSA-4QJH-9FV9-R85R Potential Timing Side-Channel Vulnerability in vLLM’s Chunk-Based Prefix Caching

This issue arises from the prefix caching mechanism, which may expose the system to a timing side-channel attack. Description When a new prompt is processed, if the PageAttention mechanism finds a matching prefix chunk, the prefill process speeds up, which is reflected in the TTFT Time to First...

Microsoft OneDrive File Picker Flaw Grants Apps Full Cloud Access — Even When Uploading Just One File

Cybersecurity researchers have discovered a security flaw in Microsoft's OneDrive File Picker that, if successfully exploited, could allow websites to access a user's entire cloud storage content, as opposed to just the files selected for upload via the tool. "This stems from overly broad OAuth...

Privacy-Preserving Prompt Personalization in Federated Learning for Multimodal Large Language Models

Prompt learning is a crucial technique for adapting pre-trained multimodal language models MLLMs to user tasks. Federated prompt personalization FPP is further developed to address data heterogeneity and local overfitting, however, it exposes personalized prompts - valuable intellectual assets - ...

Jailbreak Distillation: Renewable Safety Benchmarking

Large language models LLMs are rapidly deployed in critical applications, raising urgent needs for robust safety benchmarking. We propose Jailbreak Distillation JBDistill, a novel benchmark construction framework that "distills" jailbreak attacks into high-quality and easily-updatable safety...

Operationalizing CaMeL: Strengthening LLM Defenses for Enterprise Deployment

CaMeL Capabilities for Machine Learning introduces a capability-based sandbox to mitigate prompt injection attacks in large language model LLM agents. While effective, CaMeL assumes a trusted user prompt, omits side-channel concerns, and incurs performance tradeoffs due to its dual-LLM design. Th...

FIDO2 Authentication Does Not Work With Webpages Opened Using Microsoft Edge

Users are not able to Authenticate to a website that requires FIDO2 Authentication using a Yubikey when using Edge on VDA Devices. The users are constantly prompted to select a Smartcard device. The same users are able to Authenticate onto the same website using Chrome or Firefox inside the same...

Language of Network: a Generative Pre-Trained Model for Encrypted Traffic Comprehension

The increasing demand for privacy protection and security considerations leads to a significant rise in the proportion of encrypted network traffic. Since traffic content becomes unrecognizable after encryption, accurate analysis is challenging, making it difficult to classify applications and...

Phare: a Safety Probe for Large Language Models

Ensuring the safety of large language models LLMs is critical for responsible deployment, yet existing evaluations often prioritize performance over identifying failure modes. We introduce Phare, a multilingual diagnostic framework to probe and evaluate LLM behavior across three critical...

Semantic-Preserving Adversarial Attacks on LLMs: an Adaptive Greedy Binary Search Approach

Large Language Models LLMs increasingly rely on automatic prompt engineering in graphical user interfaces GUIs to refine user inputs and enhance response accuracy. However, the diversity of user requirements often leads to unintended misinterpretations, where automated optimizations distort...

Efficient and Stealthy Jailbreak Attacks Via Adversarial Prompt Distillation from LLMs to SLMs

Attacks on large language models LLMs in jailbreaking scenarios raise many security and ethical issues. Current jailbreak attack methods face problems such as low efficiency, high computational cost, and poor cross-model adaptability and versatility, which make it difficult to cope with the rapid...

CVE-2025-5074

A vulnerability, which was classified as critical, was found in FreeFloat FTP Server 1.0. Affected is an unknown function of the component PROMPT Command Handler. The manipulation leads to buffer overflow. It is possible to launch the attack remotely. The exploit has been disclosed to the public...

CVE-2024-45989

Monica AI Assistant desktop application v2.3.0 is vulnerable to Exposure of Sensitive Information to an Unauthorized Actor. A prompt injection allows an attacker to modify chatbot answer with an unloaded image that exfiltrates the user's sensitive chat data of the current session to a malicious...