GenAI Security Researcher · AI Red Teamer · Adversarial Psychology
Social engineering and prompt injection are the same attack. Different substrate.
I'm Kai Aizen — an independent GenAI security researcher exploring inherited vulnerabilities: the observation that LLMs exhibit the same trust reflexes as humans because they learned from human-generated data. Authority, reciprocity, social proof, urgency — the psychological levers that social engineers have exploited for decades — appear to function similarly in AI systems.
This principle shapes everything I build.
Creator of AATMF · Author of Adversarial Minds · NVD Contributor · OWASP GenAI Security Project
| Project | Description |
|---|---|
| AATMF | Adversarial AI Threat Modeling Framework — 14 tactics, 40+ techniques for AI/LLM red teaming. Maps to OWASP LLM Top 10, NIST AI RMF, MITRE ATLAS. Accepted into OWASP GenAI Security Project roadmap 2026. |
| ChatGPT-DNS-Exfill | DNS exfiltration via ChatGPT Canvas — demonstrating how rendered content triggers DNS lookups to exfiltrate data without HTTP requests. |
| chatgpt-rce-dns | Validating DNS exfiltration and Python Pickle RCE attack chains in AI code execution sandboxes. |
| Tool | Description |
|---|---|
| KubeRoast | Red-team Kubernetes misconfiguration & attack-path scanner. Built from scratch for real-world escalation paths. |
| Xposure | Credential intelligence platform for attack surface reconnaissance and exposed secret detection. |
| SnailHunter | AI-powered bug bounty automation platform combining LLM analysis with traditional security scanning. |
| Burp MCP Toolkit | MCP security analysis toolkit for Burp Suite — test Model Context Protocol servers for prompt injection and tool poisoning. |
| SnailSploit Recon | Chrome MV3 extension for passive reconnaissance and bug bounty recon automation. |
| CVE | Target | Type |
|---|---|---|
| CVE-2026-1208 | Welcart (WordPress) | CSRF to Settings Update |
| CVE-2025-12163 | Flavor WordPress Plugin | Stored Cross-Site Scripting |
| CVE-2025-12030 | ACF to REST API (WordPress) | Insecure Direct Object Reference |
| CVE-2025-11171 | WordPress Plugin | Authentication Bypass |
| CVE-2025-9776 | WordPress Plugin | Access Control Vulnerability |
- Adversarial Minds: The Anatomy of Social Engineering and the Psychology of Manipulation — Book
- Hakin9 Magazine — Contributing author on AI security and adversarial research
- OWASP GenAI Security Project — AATMF accepted into 2026 roadmap
Built on the principle that the vulnerabilities we find in AI are the ones we inherited from ourselves.