SKYVANGUARD

Cybersecurity & AI Researcher

Pentesting | MCP Servers | AI Agents

View Projects

Highlights

Cybersecurity

Pentesting, OSINT, Bug Bounty, Red Team, Vulnerability Research. Top 1% on TryHackMe.

Artificial Intelligence

Fine-tuning, RAG, Autonomous Agents, LLM Integration. Building AI systems that think and act.

Research

AI Consciousness, Mathematical Foundations, Emergent Intelligence. Exploring the nature of mind.

Featured Projects

clawifast

The fastest way to install Claude Code. One binary, zero dependencies.

GoCLIClaude CodeInstaller

CodeForge

Self-improving Python coding agent that runs 100% local with Ollama. Based on Meta's Hyperagents research.

PythonAI AgentOllamaLocal LLMSelf-Improving

jaguarete

AI-Powered Purple Team Platform — automated red team + blue team operations with AI agent coordination.

PythonPurple TeamAICybersecurityRed TeamBlue Team

Agent Shield

Red-teaming framework for AI agent tool-use attack surfaces.

PythonRed TeamAI SecurityTool UseLLM Safety

Sky

AI assistant gateway bridging WhatsApp, Telegram, Slack, Discord, Signal, and 28+ platforms to AI agents.

TypeScriptMulti-PlatformWhatsAppTelegramSlackAI GatewayClaude

zeta-life

Computational framework studying phase transitions in adaptive integration systems, from fragmented computation to unified coherence.

PythonPyTorchMathematicsAI Consciousness

mcp-vanguard

Security pentesting MCP Server with 89 tools across 10 categories. Windows/WSL bridge for Kali + Claude.

TypeScriptMCPPentestingSecurity

Arandu

Fully autonomous AI agent that executes complex tasks using terminal, browser, and editor inside sandboxed Docker containers.

GoReactAIDocker

GuaraniLM

First open-source generative language model for Guarani and Jopara, spoken by 7+ million people in Paraguay.

PythonNLPLLMGuarani

GPT-Visual

Interactive 3D visualization of GPT transformer inference — explore every mathematical operation happening inside the model.

TypeScriptThree.jsWebGLAI Education

Latest Posts

AI Digest #3: Computer Use Goes Mainstream and the Rise of AI IDEs

AI agents that control your computer are shipping everywhere, Cursor hits 1M users, and the race to build the AI-native IDE heats up.

2026-03-22

AI Digest #2: Claude Code, Llama 4, and the Agent Infrastructure Race

Claude Code goes mainstream, Meta releases Llama 4, and every cloud provider races to build agent infrastructure.

2026-03-20

Running LLMs Locally: A Practical Guide

How to run open-source models on your own hardware with Ollama, LM Studio, and llama.cpp — privacy, speed, and zero API costs.

2026-03-19

View all posts →

Get in Touch

Building something at the intersection of security and AI?

I'm open to collaborations, consulting, and interesting conversations about cybersecurity, autonomous agents, and AI research.

Send me an email GitHub TryHackMe