- Display token metrics (input/output/cache) per response in Discussion Hub.
- Add total Discussion Token usage in the panel header.
- Implement 'Compress' feature to intelligently summarize and replace exhausted discussion histories using an AI subagent.