Guardrails

Biblioteca NVIDIA NeMo Guardrails ÚLTIMA VERSIÓN / VERSIÓN DE DESARROLLO : La rama develop sigue lo último en desarrollo. La última versión publicada es la 0.24.1. ✨✨✨ 📌 La documentación oficial de la biblioteca NeMo Guardrails está disponible endocs.nvidia.com/nemo/guardrails. ✨✨✨ La biblioteca...

NemoClaw

NVIDIA NemoClaw：OpenShell 中沙箱化 AI 智能体的参考栈 NVIDIA NemoClaw 是一个开源参考栈，用于在 NVIDIA OpenShell 沙箱中更安全地运行受支持的 AI 智能体。 它通过 NemoClaw CLI 及其特定于智能体的别名，提供引导式入门、托管推理、网络策略、托管集成、由 OpenShell 支持的状态持久化以及生命周期操作。 支持的智能体： OpenClaw（默认） Hermes LangChain Deep Agents Code 有关功能、架构、安全控制以及完整功能列表，请参阅 NemoClaw 文档。 快速开始...

sass-king

SASS King 从受控内核到生产审计的 NVIDIA SASS 逆向工程。 文章 1 · 文章 2 · 知识库 · 模式库 · SM120 指令词汇表 · 编码笔记 · 从这里开始 · 项目结构 · 张量核心章节 · 贡献指南 SASS King 是一个针对 NVIDIA SASS 的系统性逆向工程项目，SASS 是编译后的 CUDA 二进制文件中发出的原生 GPU 指令集。该项目从 SM120 / SM120a 消费级 Blackwell 硬件开始，并逐步向完整的跨架构 ISA 和模式库扩展。 目标是务实的：帮助内核工程师打开一个 SASS...

CVE-2019-5700

CVE-2019-5700 遗憾的是，我没有为这个漏洞取一个时髦又巧妙的名字。这个 bug 利用起来相当简单。我实际上把它搁置了相当长一段时间，直到我意识到它可能影响的设备比我最初想到的要多。 影响范围： 此漏洞影响启动 Android 启动映像的 Nvidia Tegra t114-t210 SoC，但有一个重要的区别需要说明： 基于 t114 的设备受影响严重得多，因为其统一引导加载程序同时加载 TOS（TLK - Nvidia 的安全监视器），这意味着引导加载程序已经在安全上下文中运行…… 在 t124+ 设备上，Nvidia 现在使用 nvtboot...

scallion

2020 Project Update Scallion was a hobby project to learn OpenCL and RSA. We are not currently maintaining scallion. It has bitrotted in two important ways: 1. Tor has moved on to v3 Onion addresses, which scallion does not support. Old v2 addresses will stop working in October of 2021. The new...

SkillSpector

SkillSpector Security scanner for AI agent skills. Detect vulnerabilities, malicious patterns, and security risks before installing agent skills. Overview AI agent skills used by Claude Code, Codex CLI, Gemini CLI, etc. execute with implicit trust and minimal vetting. In the 31,132-skill analyzed...

Flowise - NVIDIA NIM Endpoints Missing Authentication

Flowise is a drag & drop user interface to build a customized large language model flow. Prior to version 3.0.13, the NVIDIA NIM router /api/v1/nvidia-nim/ is whitelisted in the global authentication middleware, allowing unauthenticated access to privileged container management and token generati...

NVIDIA Triton Inference Server <= 26.02 - Authentication Bypass

NVIDIA Triton Inference Server contains an authentication bypass vulnerability, letting attackers bypass authentication and potentially execute code, escalate privileges, tamper data, cause denial of service, or disclose information, exploit requires no special conditions. id: CVE-2026-24207 info...

EUVD-2026-94047

NVIDIA TensorRT contains a vulnerability where an attacker can cause an out of bounds read. A successful exploit of this vulnerability may lead to denial of service...

EUVD-2026-94049

NVIDIA Model-Optimizer contains a vulnerability where an attacker may cause deserialization of untrusted data. A successful exploit of this vulnerability might lead to code execution, data tampering, denial of service, and information disclosure...

CVE-2026-65122

NVIDIA TensorRT contains a vulnerability where an attacker can cause an out of bounds read. A successful exploit of this vulnerability may lead to denial of service...

CVE-2026-65142

NVIDIA Model-Optimizer contains a vulnerability where an attacker may cause deserialization of untrusted data. A successful exploit of this vulnerability might lead to code execution, data tampering, denial of service, and information disclosure...

CVE-2021-1056

CVE-2021-1056 CVE-2021-1056 是我向 NVIDIA PSIRT 提交的一个漏洞。个人认为，在多租户 HPC 集群（尤其是云机器学习平台）中，它可能导致较高的安全风险。 本仓库简单地演示了该漏洞在由 nvidia-container-runtime 创建的 GPU 容器中的表现。 工作原理 通过创建特定的字符设备文件，GPU 容器（由 nvidia-container-runtime 创建的容器）中的攻击者能够访问宿主机上所有的 GPU 设备。 该漏洞同样适用于在 Kubernetes 集群上由 k8s-device-plugin 创建的 GPU Pod。 前提条...

CVE-2026-65142

NVIDIA Model-Optimizer contains a vulnerability where an attacker may cause deserialization of untrusted data. A successful exploit of this vulnerability might lead to code execution, data tampering, denial of service, and information disclosure...

CVE-2026-65142

NVIDIA Model-Optimizer contains a vulnerability where an attacker may cause deserialization of untrusted data. A successful exploit of this vulnerability might lead to code execution, data tampering, denial of service, and information disclosure...

CVE-2026-65142

The NVIDIA Model-Optimizer software is vulnerable to the deserialization of untrusted data (CWE-502). A successful exploit could allow an attacker to achieve code execution , data tampering , denial of service , and information disclosure . No specific affected versions, remediation steps, or exp...

CVE-2026-65122

The NVIDIA TensorRT library is susceptible to an out-of-bounds read vulnerability (CWE-125). A successful exploitation of this flaw can result in a denial of service (DoS). The provided documents do not specify the affected versions, specific components, root causes, or available remediation step...

CVE-2026-65122

NVIDIA TensorRT contains a vulnerability where an attacker can cause an out of bounds read. A successful exploit of this vulnerability may lead to denial of service...

colabcat

Colabcat Run Hashcat on Google Colab with session restore capabilities with Google Drive. Usage Go to the link below to open a copy of the colabcat.ipynb file in Google Colab: https://colab.research.google.com/github/someshkar/colabcat/blob/master/colabcat.ipynb Click on Runtime, Change runtime...

hashcrack

hashcrack desempaqueta algunos tipos de hash, elige opciones sensatas e invoca hashcat Controladores NVIDIA aquí: http://www.nvidia.co.uk/Download/index.aspx?lang=en-uk Correo electrónico alternativo: [email protected] si necesitas soporte No estoy seguro de si alguien en NCC se hará cargo d...