Independent Security Researcher · Offensive Security Engineer · Tool Builder
ROT Independent Security Research Lab · Chennai, India

• LLM Security & AI Red Teaming — adversarial prompt evolution, guardrail bypass, differential testing across LLM providers
• Protocol Fuzzing — coverage-guided and mutation-based fuzzing of network protocol implementations
• Offensive Security Automation — high-performance exploit research infrastructure, adversary simulation systems
• Language Engineering — compiler design, bytecode VMs, security-oriented programming languages
• WebSocket Security — protocol-level vulnerability discovery, payload mutation, headless DOM verification

Basilisk: An Evolutionary AI Red-Teaming Framework for Systematic Security Evaluation of Large Language Models
Regaan · ROT Independent Security Research Lab · March 2026
📄 DOI: 10.5281/zenodo.18909538 · Indexed: OpenAIRE · License: CC BY 4.0

Introduces Smart Prompt Evolution (SPE-NL) — a genetic algorithm treating adversarial prompts as organisms under selection pressure. 29 attack modules across 8 OWASP LLM Top 10 categories. 92% improvement in attack success rate over static payload libraries at generation 5.

Open-source framework for systematic adversarial security evaluation of Large Language Models using evolutionary computation.

• SPE-NL engine — genetic algorithm with 10 mutation operators, 5 crossover strategies, multi-signal fitness function
• 29 attack modules — prompt injection, system prompt extraction, data exfiltration, tool abuse, guardrail bypass, DoS, multi-turn manipulation, RAG attacks
• 100+ LLM providers via LiteLLM abstraction layer
• Differential testing — simultaneous cross-provider behavioral divergence analysis
• Non-destructive posture assessment — production-safe guardrail grading (A+ to F)
• SARIF 2.1.0 output for GitHub Code Scanning integration
• Cryptographic audit trail — SHA-256 chain integrity, auto secret redaction

pip install basilisk-ai
basilisk scan --target https://your-llm-endpoint.com --mode standard

Enterprise-grade distributed WebSocket security testing framework. 900+ repository clones · 120+ active developers.

• Genetic payload mutation engine with 22k+ corpus
• Headless DOM verification via Playwright
• 22 integrated web pentesting modules (XSS, SQLi, SSRF, OAST)
• Real-time MitM interceptor
• Distributed fuzzing architecture in Go

Coverage-guided mutation fuzzer for network protocol implementations. 350k+ executions/sec in distributed environments.

LLM fine-tuned on 1,400+ exploit samples and CVE data for automated proof-of-concept generation and shellcode research.

Modern WAF bypass tamper scripts targeting Cloudflare, AWS WAF, and Azure WAF using 2025 evasion techniques.

LLVM-based statically typed systems language engineered for offensive security research and high-performance exploit primitives.

Interpreted language with native async/await, advanced generators, and deterministic control flow.

• 📜 Published Researcher — Basilisk paper, Zenodo DOI: 10.5281/zenodo.18909538 (2026)
• 🎓 CEH — Certified Ethical Hacker, EC-Council (2025–2026)
• 🏆 OffSec Echo — Completed 9 weeks: Forensics, Cloud, Malware Analysis
• 🎤 Guest Lecturer — MIT App Development, DDGDVC Chennai (2026)
• 🎤 Guest Speaker — Offensive R&D Workshop, Mohamed Sathak Polytechnic College (2026)
• 🎤 Workshop Instructor — Modern Cyber Attack Engineering, eHackify (2025)

All security research and tooling is conducted in authorized environments only.
ROT Independent Security Research Lab · © 2026 Regaan