Agents Rule of Two: A Practical Approach to AI Agent Security | Not Hacker News!