Why do AI agents repeat mistakes despite saving past lessons? Explore how retrieval, hooks, context engineering, and memory selection shape agent behavior.