AI Deception Layer for - containing adversarial context designed to redirect or confuse an LLM agent reading your own infrastructure - View it on GitHub