Related papers: MCP-SandboxScan: WASM-based Secure Execution and R…

Large Language Models (LLMs) are increasingly deployed in sensitive domains including healthcare, legal services, and confidential communications, where privacy is paramount. This paper introduces Whisper Leak, a side-channel attack that…

Large language models are increasingly used as orchestrators of external tools via the Model Context Protocol (MCP), but MCP is built for software services with megabytes of memory and does not descend to the microcontrollers that dominate…

Current cyber-physical systems (CPS) are expected to accomplish complex tasks. To achieve this goal, high performance, but unverified controllers (e.g. deep neural network, black-box controllers from third parties) are applied, which makes…

Every attack begins with gathering information about the target. The entry point for network breaches are often vulnerabilities in internet facing websites, which often rely on an off-the-shelf Content Management System (CMS). Bot networks…

Large Language Model (LLM) systems are inherently compositional, with individual LLM serving as the core foundation with additional layers of objects such as plugins, sandbox, and so on. Along with the great potential, there are also…

Agentic AI systems and Physical or Embodied AI systems have been two key research verticals at the forefront of Artificial Intelligence and Robotics, with Model Context Protocol (MCP) increasingly becoming a key component and enabler of…

Formal methods for guaranteeing that a protocol satisfies a cryptographic security definition have advanced substantially, but such methods are still labor intensive and the need remains for an automated tool that can positively identify an…

Software vulnerabilities remain a significant risk factor in achieving security objectives within software development organizations. This is especially true where either proprietary or open-source software (OSS) is included in the…

We introduce a new mechanism to securely extend systems data collection software with potentially untrusted third-party code. Unlike existing tools which run extension modules or plugins directly inside the monitored endpoint (the guest),…

Large language models (LLMs) have shown strong capabilities in multi-step decision-making, planning and actions, and are increasingly integrated into various real-world applications. It is concerning whether their strong problem-solving…

Tool-using LLM agents must act on untrusted webpages, emails, files, and API outputs while issuing privileged tool calls. Existing defenses often mediate trust at the granularity of an entire tool invocation, forcing a brittle choice in…

Malicious software threats and their detection have been gaining importance as a subdomain of information security due to the expansion of ICT applications in daily settings. A major challenge in designing and developing anti-malware…

LLM agents process trusted instructions, retrieved records, and tool observations through a common generative channel. This conflates data flow with authority: an untrusted string can affect a secret-bearing response or an action proposal…

Prompt leakage poses a compelling security and privacy threat in LLM applications. Leakage of system prompts may compromise intellectual property, and act as adversarial reconnaissance for an attacker. A systematic evaluation of prompt…

Large Language Model (LLM) agents increasingly act through external tools, making their safety contingent on tool-call workflows rather than text generation alone. While recent benchmarks evaluate agents across diverse environments and risk…

Large language models (LLMs) are being increasingly integrated into practical hardware and firmware development pipelines for code generation. Existing studies have primarily focused on evaluating the functional correctness of LLM-generated…

As modern systems increasingly rely on GPUs for computationally intensive tasks such as machine learning acceleration, ensuring the integrity of GPU computation has become critically important. Recent studies have shown that GPU kernels are…

Search agents connect LLMs to the Internet, enabling them to access broader and more up-to-date information. However, this also introduces a new threat surface: unreliable search results can mislead agents into producing unsafe outputs.…

Large-scale vision-language models, especially CLIP, have demonstrated remarkable performance across diverse downstream tasks. Soft prompts, as carefully crafted modules that efficiently adapt vision-language models to specific tasks,…

LLM agents are increasingly deployed as executable systems that use tools, modify workspaces, and produce concrete artifacts. In such workflows, performance depends not only on the base model, but also on the harness: the system layer that…