Improper Neutralization of Special Elements Used in a Template Engine

Overview Affected versions of this package are vulnerable to Improper Neutralization of Special Elements Used in a Template Engine via the formatblockparametertemplatefromworkflowruncontext function in the block.py file. An attacker can access sensitive information by exploiting the Jinja runtime...

Can In-Context Reinforcement Learning Recover from Reward Poisoning Attacks?

We study the corruption-robustness of in-context reinforcement learning ICRL, focusing on the Decision-Pretrained Transformer DPT, Lee et al., 2023. To address the challenge of reward poisoning attacks targeting the DPT, we propose a novel adversarial training framework, called Adversarially...

Exploit for Privilege Context Switching Error in Canonical Ubuntu_Linux

PoC exploit for CVE-2021-3493, an Ubuntu OverlayFS Local Privesc vulnerability. The target is the Linux kernel, specifically the overlayfs file system, which did not properly validate the application of file system capabilities with respect to user namespaces. A local attacker could use this to...

CVE-2025-38002 io_uring/fdinfo: grab ctx->uring_lock around io_uring_show_fdinfo()

In the Linux kernel, the following vulnerability has been resolved: iouring/fdinfo: grab ctx-uringlock around iouringshowfdinfo Not everything requires locking in there, which is why the 'haslock' variable exists. But enough does that it's a bit unwieldy to manage. Wrap the whole thing in a...

On Automating Security Policies with Contemporary LLMs

The complexity of modern computing environments and the growing sophistication of cyber threats necessitate a more robust, adaptive, and automated approach to security enforcement. In this paper, we present a framework leveraging large language models LLMs for automating attack mitigation policy...

Hewlett Packard Enterprise Insight Remote Support processAttachmentDataStream Directory Traversal Remote Code Execution Vulnerability

This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hewlett Packard Enterprise Insight Remote Support. Authentication is not required to exploit this vulnerability. The specific flaw exists within the implementation of the processAttachmentDataStream...

TracLLM: a Generic Framework for Attributing Long Context LLMs

Long context large language models LLMs are deployed in many real-world applications such as RAG, agent, and broad LLM-integrated applications. Given an instruction and a long context e.g., documents, PDF files, webpages, a long context LLM can generate an output grounded in the provided context,...

2BrightSparks SyncBackFree Link Following Local Privilege Escalation Vulnerability

This vulnerability allows local attackers to escalate privileges on affected installations of 2BrightSparks SyncBackFree. An attacker must first obtain the ability to execute low-privileged code on the target system in order to exploit this vulnerability. User interaction on the part of an...

PT-2025-30774

Name of the Vulnerable Software and Affected Versions Linux kernel affected versions not specified Description The OP-TEE driver registers the notif callback function for FF-A notifications. This function is called in an atomic context, which can lead to errors when processing asynchronous...

PT-2025-38448

Name of the Vulnerable Software and Affected Versions Linux kernel affected versions not specified Description The extent writepage function in the Linux kernel incorrectly handles error conditions by setting the PageError flag whenever an error occurs and then checking for this flag to determine...

Chances and Challenges of the Model Context Protocol in Digital Forensics and Incident Response

Large language models hold considerable promise for supporting forensic investigations, but their widespread adoption is hindered by a lack of transparency, explainability, and reproducibility. This paper explores how the emerging Model Context Protocol can address these challenges and support th...

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...

Markdownify MCP Server 安全漏洞

Markdownify MCP Server is a Model Context Protocol server for converting almost any content to Markdown by Zach Caceres, an individual developer in the United States. A security vulnerability exists in Markdownify MCP Server that stems from the Markdownify.get function that could lead to...

Key Takeaways from the Take Command Summit 2025: Customer Panel on Future-Proofing VM Programs

One of the most actionable sessions at the Take Command 2025 Virtual Cybersecurity Summit came directly from the field. In a panel hosted by Aniket Menon, VP of Product Management at Rapid7, security leaders from Cross Financial Corp, Phibro Animal Health Corporation, and Miltenyi Biotec shared h...

aws-mcp-server 操作系统命令注入漏洞

aws-mcp-server is a lightweight service by Alexei Ledenev Personal Developer that enables AI assistants to execute AWS CLI commands in a secure containerized environment via the Model Context Protocol MCP. A security vulnerability exists in aws-mcp-server that stems from command injection and cou...

VulBinLLM: LLM-Powered Vulnerability Detection for Stripped Binaries

Recognizing vulnerabilities in stripped binary files presents a significant challenge in software security. Although some progress has been made in generating human-readable information from decompiled binary files with Large Language Models LLMs, effectively and scalably detecting vulnerabilitie...

Cross-Site Scripting (XSS)

dotnetnuke.core is vulnerable to cross-site scripting XSS. The vulnerability is due to uploaded SVG files containing scripts that, when rendered inline. It allows an attacker to execute malicious scripts in the context of the user’s browser...

CVE-2025-48382 Fess has Insecure Temporary File Permissions

Fess is a deployable Enterprise Search Server. Prior to version 14.19.2, the createTempFile method in org.codelibs.fess.helper.SystemHelper creates temporary files without explicitly setting restrictive permissions. This could lead to potential information disclosure, allowing unauthorized local...

What Really Matters in Many-Shot Attacks? an Empirical Study of Long-Context Vulnerabilities in LLMs

We investigate long-context vulnerabilities in Large Language Models LLMs through Many-Shot Jailbreaking MSJ. Our experiments utilize context length of up to 128K tokens. Through comprehensive analysis with various many-shot attack settings with different instruction styles, shot density, topic,...

CVE-2025-24784

kubewarden-controller is a Kubernetes controller that allows you to dynamically register Kubewarden admission policies. The policy group feature, added to by the 1.17.0 release. By being namespaced, the AdmissionPolicyGroup has a well constrained impact on cluster resources. Hence, it’s considere...