wasmtime 安全漏洞

wasmtime is a lightweight WebAssembly runtime open-sourced by the Bytecode Alliance. A security vulnerability exists in wasmtime version 38.0.0 through versions prior to 38.0.3, which stems from a flaw in the host-to-wasm springboard implementation associated with the component model, which could...

Jailbreak Mimicry: Automated Discovery of Narrative-Based Jailbreaks for Large Language Models

Large language models LLMs remain vulnerable to sophisticated prompt engineering attacks that exploit contextual framing to bypass safety mechanisms, posing significant risks in cybersecurity applications. We introduce Jailbreak Mimicry, a systematic methodology for training compact attacker mode...

CVE-2021-43768

CVE-2021-43768 affects Malwarebytes For Teams up to v1.0.990; the issue is a privilege escalation via the COM interface running in mbamservice.exe. It is fixed in v1.0.1003 and later. Affected component is the COM interface, which can be leveraged to elevate privileges within the host. Remediatio...

Securing AI Agent Execution

Large Language Models LLMs have evolved into AI agents that interact with external tools and environments to perform complex tasks. The Model Context Protocol MCP has become the de facto standard for connecting agents with such resources, but security has lagged behind: thousands of MCP servers...

PT-2025-43631

D-Link DIR600L Ax FW116WWb01 was discovered to contain a buffer overflow via the curTime parameter in the function formAdvFirewall...

The Trojan Example: Jailbreaking LLMs through Template Filling and Unsafety Reasoning

Large Language Models LLMs have advanced rapidly and now encode extensive world knowledge. Despite safety fine-tuning, however, they remain susceptible to adversarial prompts that elicit harmful content. Existing jailbreak techniques fall into two categories: white-box methods e.g., gradient-base...

NewStart CGSL MAIN 7.02 : pytorch Vulnerability (NS-SA-2025-0250)

The remote NewStart CGSL host, running version MAIN 7.02, has pytorch packages installed that are affected by a vulnerability: - PyTorch is a Python package that provides tensor computation with strong GPU acceleration and deep neural networks built on a tape-based autograd system. In version 2.5...

MINI-G968-G5V9-34XH

Bulletin has no description...

Shadow Escape 0-Click Attack in AI Assistants Puts Trillions of Records at Risk

Operant AI reveals Shadow Escape, a zero-click attack using the MCP flaw in ChatGPT, Gemini, and Claude to secretly steal trillions of SSNs and financial data. Traditional security is blind to this new AI threat...

CVE-2025-49923

Improper Neutralization of Input During Web Page Generation 'Cross-site Scripting' vulnerability in Craig Hewitt Seriously Simple Podcasting seriously-simple-podcasting allows DOM-Based XSS.This issue affects Seriously Simple Podcasting: from n/a through = 3.11.1...

CVE-2025-53702

Vilar VS-IPC1002 IP cameras are affected by CVE-2025-53702 affecting the DoS vector via crafted requests to /cgi-bin/action on the same local network. An unauthenticated attacker can render the device unresponsive, requiring a manual restart. Only version 1.1.0.18 was tested; other versions might...

Apple may have to open its walled garden to outside app stores

The UK’s Competition and Markets Authority CMA ruled that both Google and Apple have a "strategic market status." Basically, they have a monopoly over their respective mobile platforms. As a result, Apple may soon be required to allow rival app stores on iPhones—a major shift for the smartphone...

CVE-2025-11915

Connection desynchronization between an HTTP proxy and the model backend. The fixes were rolled out for all proxies in front of impacted models by 2025-09-28. Users do not need to take any action...

REx86: A Local Large Language Model for Assisting in X86 Assembly Reverse Engineering

Reverse engineering RE of x86 binaries is indispensable for malware and firmware analysis, but remains slow due to stripped metadata and adversarial obfuscation. Large Language Models LLMs offer potential for improving RE efficiency through automated comprehension and commenting, but cloud-hosted...

Brother MFC-J4410DW Printers Cross-site Scripting (CVE-2015-1056)

Cross-site scripting XSS vulnerability in Brother MFC-J4410DW printer with firmware before L allows remote attackers to inject arbitrary web script or HTML via the url parameter to general/status.html and possibly other pages. This plugin only works with Tenable.ot. Please visit...

Enhancing Security in Deep Reinforcement Learning: A Comprehensive Survey on Adversarial Attacks and Defenses

With the wide application of deep reinforcement learning DRL techniques in complex fields such as autonomous driving, intelligent manufacturing, and smart healthcare, how to improve its security and robustness in dynamic and changeable environments has become a core issue in current research...

CVE-2025-49940

CVE-2025-49940 is a DOM-based XSS in WordPress Fusion Builder (fusion-builder) up to version 3.13.2. The root cause is improper input neutralization during web page generation, enabling script execution in the browser context. Affected software: Fusion Builder fusion-builder (WordPress plugin)

CVE-2025-49923 WordPress Seriously Simple Podcasting plugin <= 3.11.1 - Cross Site Scripting (XSS) vulnerability

Improper Neutralization of Input During Web Page Generation 'Cross-site Scripting' vulnerability in Craig Hewitt Seriously Simple Podcasting seriously-simple-podcasting allows DOM-Based XSS.This issue affects Seriously Simple Podcasting: from n/a through = 3.11.1...

CVE-2025-11915

CVE-2025-11915 involves a connection desynchronization issue between an HTTP proxy and the Vertex AI model backend. Root cause: synchronization mismatch across the proxy-backend boundary. Impact is described as affecting certain third-party models, with fixes rolled out for all proxies in front o...

EUVD-2025-35358

Connection desynchronization between an HTTP proxy and the model backend. The fixes were rolled out for all proxies in front of impacted models by 2025-09-28. Users do not need to take any action...