Security is a pillar of sovereignty, so I built a hacking lab to learn how I get hacked
Data sovereignty is usually sold as hiding, but real control starts with understanding how your data is actually taken. To learn that honest half of the story, I built a fully isolated hacking lab on my own desktop, in a single afternoon, together with Claude Code. It is a small copy of the real world that touches nothing real: one private network with no route to my home or the internet, an attacker machine with the full Kali toolkit, a set of deliberately vulnerable Linux and web targets, and a miniature Windows domain. Every target resets from a snapshot, and every "secret" inside, the bank statement, the crypto seed, the reused password, is a planted fake. You practice the attack so you earn the defense. What makes the lab more than a sandbox is its tutor: an open qwen3 model running locally through Ollama, free of the usual guardrails and connected by retrieval to two private knowledge bases, the lab's own living documentation and my personal Joplin notebooks. Nothing leaves the machine. For a beginner this removes the biggest barrier, the fear of asking a stupid question, and it raises an obvious question of its own: an unlimited AI wired to offensive tools, what can go wrong? The answer is the whole design. The model has no guardrails, so the guardrails live in the isolation of the lab and in the operator. Inside a sealed room, against machines I own, that knowledge is simply education, and it is how sovereignty over your own security is earned, one brick at a time.
