#1LLM security, prompt injection, model extraction, adversarial AI, and AI red teaming tools.
Kitploit recommended

Algorithms for outlier, adversarial and drift detection
OWASP Web Security Testing Guide RAG system with ChromaDB, MCP for Claude Code

A novel adversarial attack on LLM based on the Exponentiated Gradient Descent technique.

a security scanner for custom LLM applications

CAWODOG is a proof-of-concept project demonstrating how to protect Python-based AI models deployed on offline industrial machines. Across three…

Proof-of-concept exploit for CVE-2025-32434, a pickle deserialization vulnerability in PyTorch's weights_only mode, demonstrating malicious tar file…

Burp Suite Repeater extension that automatically mutates payloads and analyzes responses to uncover path traversal, SQL injection, XSS, and other web…

My manifesto, and stories, images, and phun stuff

[ICCV 2025] Anti-Tamper Protection for Unauthorized Individual Image Generation

Code repository for CS5446 project exploring defenses against jailbreak attacks on large-language models it includes datasets, notebooks,…

Artefacts for blog post on finding CVE-2025-37899 with o3

A black-box optimization approach for protecting Large Language Model (LLM) system prompts against adversarial extraction attacks.

[NeurIPS '25] Code for Paper "IF-Guide: Influence Function-Guided Suppression of Harmful Training Data for Reducing LLM Toxicity"

Code Implementation of "Pattern Enhanced Multi-Turn Jailbreaking: Exploiting Structural Vulnerabilities in Large Language Models"

Automated framework for hijacking safety reasoning in large reasoning models via simulated reasoning traces and iterative prompt refinement to…

AI Red Teaming playground labs to run AI Red Teaming trainings including infrastructure.

Research code and experiments for defending tool-integrated LLM agents against adversarial attacks, extending Agent Security Bench with new defense…

Benchmark for evaluating safety risks of computer-using agents, with 104 realistic misuse scenarios across seven malicious categories, supporting…