Handling new worktrees more efficiently

[dghelm]

In my current workflow I will create new git worktrees and often several at a time for a new scope of work for AI agents.

At the moment, the performance here leaves my machine basically unusable for the variation of time that it takes to track these new directories. The resource competition is for disk IO and write specifically.

Has anyone come up with any good workarounds for this?

[dghelm]

Likely addressed by #654