Automatos Academy · Cross-Domain
AI Security — LLM & Agent Security
Defensive-security skills for the people who build and run LLM apps and agents. Six teaching modules take you from the LLM threat model to red-teaming your own app: OWASP Top 10 for LLM Applications (2025) as the risk backbone, MITRE ATLAS as the shared adversary map, and NIST AI 600-1 as the govern
Skills track — no exam, no gate: finish the work, ship the capstone
Facts verified 2026-07-02 against the official source
Start free — no account needed →
What's inside
Lessons grounded in primary sources · hands-on labs · branching scenario drills · a shipped capstone · videos and audio deep-dives · a source library where every claim links to the official doc.
- The LLM threat model & the OWASP Top 10
- Prompt injection & improper output handling
- Poisoning, supply chain & RAG security
- Excessive agency & agentic-system security
- The LLM gateway & red-teaming your own app
- MITRE ATLAS & NIST — the adversary map and the governance overlay
Questions
Is there an exam?
No — this is a skills track. You finish when every module's hands-on work is done and the capstone ships.
What does it cost?
Nothing. The whole academy is free — no account required, no trial, no upsell.
Do I need to be technical?
No. Modules are plain English first; every term of art is explained before it's used.
Automatos Academy is independent, free training — not affiliated with or endorsed by any certification body. Exam names are trademarks of their respective owners.