Agent Cost Optimization

Reduce an agent's cost and latency without dropping quality — measurement, prompt-cache hygiene, model routing, context diet, and batching. Use when the LLM bill spikes, when cost per task must come down, or before scaling an agent 10x.