modelaudit

ModelAudit Asegure sus modelos de IA antes del despliegue. Escáner estático que detecta código malicioso, posibles indicadores de puertas traseras y vulnerabilidades de seguridad en archivos de modelos de ML — sin cargarlos ni ejecutarlos nunca. Documentación completa | Ejemplos de uso | Formatos...

changeme

changeme Un escáner de credenciales por defecto. Acerca de changeme retoma donde los escáneres comerciales lo dejan. Se centra en detectar credenciales por defecto y puerta trasera, no necesariamente credenciales comunes. Su modo predeterminado es escanear credenciales HTTP por defecto, pero tien...

CVE-2021-44228

CVE-2021-44228 Backdoor detection for VMware view horizon This includes a canary with an optional submit. Clearly use at own risk blah blah This is designed to be run on Windows VMware Horizon connection servers. It doens't look at the UAG appliances they are Photon linux...

psad

psad - iptables 로그를 이용한 침입 탐지 소개 Port Scan Attack Detector psad는 가벼운 시스템 데몬으로, Linux iptables/ip6tables/firewalld 방화벽 코드와 함께 작동하여 포트 스캔 및 스위프, 백도어, 봇넷 명령 및 제어 통신 등 의심스러운 트래픽을 탐지하도록 설계되었습니다. 이 도구는 매우 구성 가능한 위험 임계값 세트합리적인 기본값 제공, 소스, 대상, 스캔된 포트 범위, 시작 및 종료 시간, TCP 플래그 및 해당 nmap 옵션, 역방향 DNS 정보, 이메일 ...

vscan

vscan Herramienta de escaneo de vulnerabilidades que utiliza nmap y scripts NSE para encontrar vulnerabilidades Esta herramienta aporta un valor adicional al escaneo de vulnerabilidades con nmap. Utiliza scripts NSE que pueden añadir flexibilidad en términos de detección y explotación de...

RCLocals

RCLocals Inspirado en 'Autoruns' de Sysinternals, RCLocals analiza todas las posibilidades de inicio de Linux para encontrar puertas traseras, también realiza verificación de integridad de procesos, escanea procesos con inyección de DLL y mucho más Aspectos cubiertos: ·Listar claves GPG de...

sticky_keys_hunter

스티키 키 헌터 이 POC 배시 스크립트는 스티키 키sticky keys 및 유틸맨utilman 백도어를 테스트합니다. 스크립트는 RDP 서버에 연결하고, 스티키 키 및 유틸맨 트리거를 전송한 후 결과를 스크린샷으로 저장합니다. 이 스크립트는 블랙박스 방식으로 이러한 백도어를 탐지하는 이론을 입증하기 위해 작성되었으며, 제 블로그 글 Hunting Sticky Keys Backdoors 이후로 업데이트되지 않았습니다. 하지만 @DennisMald와 @notmedic가 이 연구를 발전시켜 DEF CON 24 발표 Sticky...

TNC-Defense

TNC-Defense TNC-Defense proporciona código de investigación para detectar y desintoxicar puertas traseras en modelos de difusión de texto a imagen. El repositorio incluye actualmente pipelines de detección para Stable Diffusion v1.4, Stable Diffusion v1.5, Stable Diffusion XL y Stable Diffusion 3...

CVE-NetScalerFileSystemCheck

CVE-NetScalerFileSystemCheck Este script verifica si el Citrix Netscaler ha sido comprometido por ataques CVE-2019-19781 y recopila toda la información del sistema de archivos. Los siguientes archivos y registros serán revisados versión más reciente 1.13: Carpetas de plantilla para archivos XML...

Security Properties of Neural Networks As Decision Problems

Certifying a deployed neural network raises decision problems that the verification literature has not classified: whether the model carries a backdoor planted in its training data, whether a fault in its stored parameters can drive it into an unsafe state, whether its output leaks a private part...

Learning Normal Diffusion Dynamics for Backdoor Defense in Text-To-Image Models

Backdoor attacks pose a serious threat to the secure deployment of text-to-image T2I diffusion models. Existing defenses typically detect backdoors from specific abnormal patterns in internal representations, which may limit their generalizability with the emergence of increasingly diverse attack...

WordPress Adds Automated Plugin Reviews to Block High-Risk Updates Before Distribution

WordPress has announced it's launching an automated security review for every release of a plugin before it's distributed through the WordPress.org update API so as to analyze it for potential security issues and ensure there are no risks involved. "New plugins are reviewed before they enter the...

When Modalities Fail to Tango: Conformal Backdoor Detection in Multimodal Contrastive Learning

Backdoor attacks in multimodal contrastive learning MCL have garnered growing attention in recent years, as many downstream tasks critically depend on pre-trained MCL models. Existing detection-based defenses predominantly rely on the CLIPScore metric, under the assumption that poisoned pairs...

CLIP-Inspector: Model-Level Backdoor Detection for Prompt-Tuned CLIP Via OOD Trigger Inversion

Organisations with limited data and computational resources increasingly outsource model training to Machine Learning as a Service MLaaS providers, who adapt vision-language models VLMs such as CLIP to downstream tasks via prompt tuning rather than training from scratch. This semi-honest setting...

Detecting backdoored language models at scale

Today, we are releasing new research on detecting backdoors in open-weight language models. Our research highlights several key properties of language model backdoors, laying the groundwork for a practical scanner designed to detect backdoored models at scale and improve overall trust in AI...

Detecting backdoored language models at scale

Today, we are releasing new research on detecting backdoors in open-weight language models. Our research highlights several key properties of language model backdoors, laying the groundwork for a practical scanner designed to detect backdoored models at scale and improve overall trust in AI...

LoRA As Oracle

Backdoored and privacy-leaking deep neural networks pose a serious threat to the deployment of machine learning systems in security-critical settings. Existing defenses for backdoor detection and membership inference typically require access to clean reference models, extensive retraining, or...

Cross-LLM Generalization of Behavioral Backdoor Detection in AI Agent Supply Chains

As AI agents become integral to enterprise workflows, their reliance on shared tool libraries and pre-trained components creates significant supply chain vulnerabilities. While previous work has demonstrated behavioral backdoor detection within individual LLM architectures, the critical question ...

PoTS: Proof-Of-Training-Steps for Backdoor Detection in Large Language Models

As Large Language Models LLMs gain traction across critical domains, ensuring secure and trustworthy training processes has become a major concern. Backdoor attacks, where malicious actors inject hidden triggers into training data, are particularly insidious and difficult to detect. Existing...

Binary Diff Summarization Using Large Language Models

Security of software supply chains is necessary to ensure that software updates do not contain maliciously injected code or introduce vulnerabilities that may compromise the integrity of critical infrastructure. Verifying the integrity of software updates involves binary differential analysis...