Feedback Loop Prompting: Teach Your Agent to Critique and Retry Its Own Output

The inner-loop pattern that turns a single LLM call into an iterative refinement engine