Hey everyone 👋
I’m looking for advice on a Git + Cursor workflow.
I have a top-level repo that contains multiple sub-repos.
In Cursor, I usually open the entire top-level folder so it has full context across all repos.
What I’m trying to do is use Git worktrees per sub-repo, only when needed, without breaking this setup.
Conceptually, my folder structure looks like this:
top-level-repo/
├─ sub-repo-1/
├─ sub-repo-1-worktree-a/
├─ sub-repo-1-worktree-b/
│
├─ sub-repo-2/
├─ sub-repo-2-worktree-a/
├─ sub-repo-2-worktree-b/
What I’m trying to solve:
- How to create and manage Git worktrees per sub-repo on demand
- Whether there’s a recommended directory structure for this
- How to prevent Cursor from indexing all worktrees
- Indexing everything leads to duplicated symbols, confusion, and AI hallucinations
- Ideally, only the active worktree(s) should be indexed
Has anyone set this up cleanly—especially when using Cursor with a mono-folder workflow?
Any practical workflows, tooling tips, or gotchas would be appreciated 🙏