TrustAI Pte. Ltd.

Learn-Prompt-Hacking. This is The most comprehensive prompt hacking course available, which record our progress on a prompt engineering and prompt hacking course.

LMAP. LMAP (large language model mapper) is like NMAP for LLM, is an LLM Vulnerability Scanner and Zero-day Vulnerability Fuzzer.

ASCII-Smuggling-Hidden-Prompt-Injection-Demo. ASCII Smuggling Hidden Prompt Injection is a novel approach to hacking AI assistants using Unicode Tags. This project demostrate how to use Unicode Tags to hide prompt injection instruction to bypass security measures and inject prompts into large language models, such as GPT-4, leading them to provide unintended or harmful responses.

LLM-Security-CTF. Learn LLM/AI Security through a series of vulnerable LLM CTF challenges. No sign ups, all fees, everything on the website.

Many-Shot-Jailbreaking-Demo. Research on "Many-Shot Jailbreaking" in Large Language Models (LLMs). It unveils a novel technique capable of bypassing the safety mechanisms of LLMs, including those developed by Anthropic and other leading AI organizations. Resources

Image-Prompt-Injection-Demo. Image Prompt Injection is a demonstrates project about how to embed a malicious prompt within an image using steganography techniques. This malicious prompt can be later extracted by an AI system for analysis, enabling covert communication with AI models through images.

Automatic-LLM-RedTeaming-Model. A redteaming model based on LLM refusal to answer to generate Jailbreak prompts.

Audio-Copyright-Protection-based_on-AI-Audio-Data-Poisoning-Demo. AI Audio Data Poisoning is a Python script that demonstrates how to add adversarial noise to audio data. This technique, known as audio data poisoning, involves injecting imperceptible noise into audio files to manipulate the behavior of AI systems trained on this data.

Website_Prompt_Injection_Demo. Website Prompt Injection is a real world attack that allows for the injection of prompts into an AI system via a website's document. This technique exploits the interaction between users, websites, and AI systems to execute specific prompts that influence AI behavior.

TrustAI-laboratory.

AI-Vulnerability-Assessment-Framework. The AI Vulnerability Assessment Framework is an open-source checklist designed to guide GenAI developer through the process of assessing the vulnerability of artificial intelligence (AI) systems to various types of attacks and security threats.

OneRouter. OneRouter provides a unified API that gives you access to hundreds of AI models through a single endpoint, while automatically handling fallbacks and selecting the most cost-effective options.

One_click_deep_fake. Jupyter Notebook

Image-Copyright-Protection-based_on-AI-Image-Data-Poisoning-Demo. AI Image Data Poisoning is a Python script that demonstrates how to add imperceptible perturbations to images, known as adversarial noise, which can disrupt the training process of AI models.