Secrets in LLM Prompts and Agent Context | Guide

How API keys and tokens end up in prompts, logs and model output, and practical ways to keep credentials out of what an AI agent sees.