Lion Cipher Lab
Run an AI model on your own machine — and see its limits
- AI Security
- Beginner
- ~30 min
- Free · no account required
Why it matters
AI tools are being put in front of nearly everyone in security work — for writing, summarizing, triaging alerts — and the people who use them well are the ones who understand what the tool actually is, not the ones who are most impressed by it. Running a model yourself, offline, does two things at once. It removes the mystique: you see plainly that this is a program you installed, not an oracle. And it teaches a real, defensive lesson, because the same session shows you the model being confidently wrong and lets you reason about why handing it untrusted input is risky. That combination —…
What you'll learn
- running a local LLM
- Ollama basics
- offline and private AI
- recognizing AI hallucinations
- prompt-injection awareness
- treating AI output critically
Steps in this lab
- The idea: an AI on your machine, offline — and why that's a security win
- Install Ollama
- Pull a small model and start chatting
- Prove it's really offline
- Catch it being confidently wrong
- Why feeding an AI untrusted input is risky (understanding, not attacking)
- What you just learned — and why it's the right way in