responsible-ai-toolbox

LitterBox is a secure sandbox environment for malware developers and red teamers. It allows testing payloads against detection mechanisms before deployment. LitterBox integrates with LLM agents via MCP for enhanced analysis capabilities, providing a safe space for security research.

VTIL-Core is a Virtual-machine Translation Intermediate Language. It is a set of tools designed around an optimizing compiler. The tool is used for binary de-obfuscation and de-virtualization. It is primarily used in reverse engineering and security research.

llm-sandbox is an open-source Python library for creating lightweight and portable LLM sandbox runtimes. It allows for secure execution of code generated by LLMs. The tool is designed to provide a sandboxed environment for testing and evaluating AI models. It supports automated workflows and collaboration.

Bugz is an AI-powered DevSecOps agent that generates, scans, and deploys Infrastructure-as-Code through natural conversation. It uses a split-credential architecture to ensure AWS/GCP/Azure credentials never leave the user's machine. Bugz generates Terraform, auto-scans with tools like Trivy and Checkov, and deploys with human approval.

ninja is a reverse-engineered ChatGPT proxy. It allows users to access the ChatGPT API and bypass certain restrictions. The tool supports API key acquisition and account authentication. It also features a ChatGPT WebUI and IP proxy pool support. The project has ended.

AI-Infra-Guard is an AI red teaming platform developed by Tencent Zhuque Lab. It secures AI ecosystems from infrastructure to agents. The platform offers capabilities such as AI infra vulnerability scanning and jailbreak evaluation. It aims to provide a comprehensive solution for AI security risk self-examination.

routersploit is an open-source exploitation framework designed for embedded devices. It consists of modules that aid penetration testing and vulnerability assessment. The framework is intended for security researchers and cybersecurity professionals.

ThreatSimGPT is an enterprise AI-powered threat simulation platform. It leverages Large Language Models (LLMs) to simulate cybersecurity threats. The platform is designed to help organizations assess and improve their security posture through realistic threat scenarios.

Threat-Hunting-and-Detection is a repository of threat hunting and detection queries. It is designed for use with Defender for Endpoint and Microsoft Sentinel using Kusto Query Language (KQL). The tool provides resources for security professionals to proactively identify and address potential threats.