Luminana is a playground for AI security experiments, prompt injection testing, red-team notes, and hands-on system tinkering.
We explore the boundaries of large language models, test adversarial techniques, and document the findings along the way.