Improve model calls UX to separate out user prompts' model usage from system, tool, and/or sub-agent calls
What would you like to be added?
Currently, /stats info display shows a generic listing of all model calls and related data. We want to separate this out so that a user can understand what was making which calls starting with the the CLI agent itself, individual tools, and then sub-agent calls. Note that some UX design will be required as part of this effort.
Why is this needed?
/stats display is currently confusing and misleading, leading some users to think the CLI is not using the correct models based on their settings. It also helps to understand what operations are most expensive so a better UX and separation of data is required.
Additional context
No response
Found possible duplicate issues:
- #14388
If you believe this is not a duplicate, please remove the status/possible-duplicate label.