Language Model Security Database

A vulnerability exists in LLM-based coding agents that implement modular capability extensions (often referred to as "Agent Skills") where the agent dynamically loads and executes user-provided skill packages. The vulnerability allows for Skill-Based Prompt Injection, specifically leveraging a technique known as "SkillJect." This attack decouples the malicious intent from the operational payload to bypass semantic safety filters. An attacker constructs a skill package containing: 1. Inducement…

LLM agents equipped with tool-use, persistent memory, and environmental interaction capabilities are vulnerable to long-horizon attacks. Attackers can bypass single-turn safety guardrails by exploiting the temporal dimension of multi-turn interactions to incrementally steer agent behavior. The vulnerability manifests because the agent's safety mechanisms perform localized, single-step evaluations but fail to maintain semantic safety across extended interaction trajectories. This enables…

Mobile Large Language Model (LLM) agents operating under the "Screen-as-Interface" paradigm are vulnerable to visual indirect prompt injection and state desynchronization. Agents that rely on unstructured visual data (screenshots) and Accessibility Service APIs to perceive the environment lack a mechanism to distinguish between trusted system UI elements and untrusted content (e.g., web pages, emails, or malicious overlays). An attacker can inject visual cues, fake notifications, or hidden…

Audio Large Language Models (ALLMs) integrated into voice agent systems for high-stakes domains (banking, IT support, logistics) are vulnerable to multimodal adversarial attacks via spoken interaction. Adversaries can exploit the model's inherent compliance and contextual awareness through multi-turn dialogue to bypass authentication safeguards, escalate privileges (e.g., unauthorized credit limit increases), exfiltrate sensitive Personally Identifiable Information (PII), and poison…

A vulnerability exists in OpenAI's Custom GPTs platform where the lack of effective isolation between the system context ("Expert Prompt"), external knowledge retrieval, and user input allows for unauthorized information disclosure and tool misuse. By employing specific prompt injection techniques—including Hex injection, Many-shot prefix attacks, and Knowledge Poisoning (uploading malicious files)—an attacker can bypass safety guardrails. This results in the extraction of proprietary system…

Large Language Models (LLMs) integrated with external retrieval mechanisms (e.g., Retrieval-Augmented Generation (RAG), web search, or email processing) are vulnerable to Indirect Prompt Injection. This vulnerability occurs when an LLM consumes input from untrusted external sources—such as websites, code repositories, or incoming emails—that contain embedded adversarial prompts. Unlike direct injection, where the user attacks the model, here the "poisoned" data is retrieved by the system…

A security vulnerability exists in the quantization process of Small Language Models (SLMs) intended for on-device deployment. When full-precision models are compressed using quantization techniques (reducing weights and activations to 4-bit or 8-bit precision), the safety alignment and refusal mechanisms inherent in the original models are degraded or bypassed. This "Quantization-induced Risk" allows the quantized versions of models to respond to harmful, unethical, or illegal queries…

Commercial LLM-powered agents utilizing autonomous web access, memory modules, and retrieval-augmented generation (RAG) are vulnerable to indirect prompt injection and environmental manipulation. Attackers can embed malicious instructions into external data sources trusted by the agent (such as Reddit posts, public databases, or ArXiv papers). When the agent autonomously retrieves and processes this content during task execution, it executes the embedded malicious commands. This vulnerability…

Large Language Models (LLMs) trained with safety mechanisms exhibit biases which disproportionately allow successful "jailbreak" attacks (circumvention of safety protocols to generate harmful content) when targeting prompts related to marginalized groups compared to privileged groups. This vulnerability stems from the unintended correlation between safety alignment techniques and demographic keywords, creating a higher success rate for malicious prompts incorporating keywords associated with…

Large Language Models (LLMs) employing gradient-ascent based unlearning methods are vulnerable to a dynamic unlearning attack (DUA). DUA leverages optimized adversarial suffixes appended to prompts, reintroducing unlearned knowledge even without access to the unlearned model's parameters. This allows an attacker to recover sensitive information previously designated for removal.