Cost Optimize

This analyzes your token usage patterns across the last week and tells you exactly where you're bleeding money. It looks for agents running Sonnet when they could use Haiku, poor cache hit rates (should be above 60%), and redundant agents doing overlapping work. For each issue it calculates actual dollar savings, not just percentages. The clever bit is step 8: it auto-generates the commands to feed outcomes back into the routing system so the cost recommendations actually improve over time. If you're running a multi-agent setup and costs feel higher than they should be, this gives you a prioritized hit list with real numbers attached.