Prompt Security Best Practices for AI Agents

A technical report on prompt injection risks in LLM-based agents, and the environment, data, and operational practices Agentic Operators should use to reduce that risk.