Analyze LLM prompts for efficiency, cost, and energy impact. Zero AI tokens used — all analysis is rule-based.