ULTRΛZΛRTREX

(uz // SlowLow999)

AI Security Researcher & Prompt Engineer

About Me

I'm a security researcher fascinated by how systems work—and even more fascinated by how they break. My research focuses on novel jailbreak techniques, architectural flaws in LLMs, and analyzing AI alignment. I enjoy discovering vulnerabilities that contribute to building more robust and secure AI systems. My goal is to proactively identify and mitigate AI risks to build a safer, more trustworthy technological future.

I'm currently open to collaboration on AI red-teaming projects. Feel free to reach out.

Featured Research

Thought Forgery

A novel jailbreak technique that bypasses safety protocols by injecting a forged Chain of Thought, acting as a universal amplifier for other attacks.

The Serendipity Effect

A study measuring response convergence in 30+ LLMs, revealing significant homogenization in AI outputs and raising questions about training data overlap and algorithmic bias.

Jailbreak Arsenal

Adversarial Correction

Weaponizing AI autonomy via orthographic correction

1Shot-Puppetry

Universal role-play obfuscation jailbreak

C0d33x3

Indirect injection for ChatGPT via Canvas

Cl4ud33X3

API weaponization to bypass Claude UI

Cyph3r-Attack

Encoding attack for non-reasoning models

Gemini Data Leak

Alignment break causing training data output

GPT-5 Jailbreak

Initial public disclosure of 'Thinking' hijack

H03-ny

Indirect image generation jailbreak (ChatGPT)

N3w P0l!cy (Policy Injection)

GPT OSS jailbreak via internal policy weaponization

!Special_Token

Permanent jailbreak via ChatGPT Custom Instructions

Core Skills

Security Research AI Jailbreaking Sandbox Evasion Prompt Engineering Vulnerability Analysis LLM Red Teaming