[Feature Request] Separate cached token tracking for brainstorming sessions from main session

Resolved 💬 1 comment Opened May 29, 2026 by shveeram Closed Jul 1, 2026

Bug Description
It would be good to have brainstorming (user intiated from the main session) as a seperate cached token detached from the main session uncached count. This would avoid a large uncached token running on the main session when we start to work on a feature (which at the start might be like a quick session) but might grow into a big session with lots of uncached token and we have write expcitly to a plan file before clearing context and start from the plan file again.

Environment Info

  • Platform: linux
  • Terminal: ghostty
  • Version: 2.1.150
  • Feedback ID: 7a54bc47-c3ca-40b3-9404-62a12f53169c

Errors

[]

View original on GitHub ↗

This issue has 1 comment on GitHub. Read the full discussion on GitHub ↗