ai

This section will be for AI security notes, prompt injection, agent behavior, unsafe tool use, memory, model-connected applications, and whatever weirdness shows up when trust boundaries get blurry.