-
Notifications
You must be signed in to change notification settings - Fork 53
Description
Feature Summary
Being able to run multiple agents inside the same worktree that should be able to interfere / work on the same code - to re-initialize chats after for example running a long thread and then wanting to review with a fresh 100% context chat.
User Feedback:
"in the existing setup I am stuck, whereas with a new tab and model (or even same model) I can start fresh to continue.
other example would be code review, let's say I have just spent an hour building out a feature with claude code, then I want to do a code review before continuing, I don't want to interupt my current flow to have sonnet continue our chat to do a code review, because it could use a whole context window and now our original task is forgotten."
Problem or Use Case
users want to review the changes of one agent with another agent that has fresh context and sees the changes the other agent has made to the code already
Proposed Solution
tabs within one task
Mark from Discord has an interesting approach to this https://github.com/nof0xgiven/emdash
"Instead I can open up another tab with any model "run our test suite" or "review all the code changes" this is a fresh context window and model that can output a review, then I can go back to original that still has all the context of it's work and continue with the fixes from the review...."
Alternatives Considered
No response
Additional Context
No response