I'd love to have more visibility into the token usage. I have an agent that has considerable variance in tokens the first LLM request. It usually sits around 3300 tokens in the first request, but some calls start out at over 6000 tokens. The agent does not have a knowledge base. If we could have more visibility into what's being sent to the LLM with each request it would be really helpful and enable us to optimize. https://cdn.discordapp.com/attachments/1488289659611250791/1488289660156379327/Screenshot_2026-03-30_at_3.28.41_PM.png?ex=69cc3d42&is=69caebc2&hm=169581686d8b2581848a29cd1804777490f48618681b9227eaf9f76699d4a591&