It looks like worktrees are implicitly created for sessions.

It looks like worktrees are implicitly created for sessions. I think that there should be some prompt, at least the first time, to ask what branch the worktree should be created from.

For example, I was working my local dotnet/roslyn repo, which has both `origin` and `upstream` remotes. My `origin/master` branch is from ~2021. I think `upstream` has no `master` branch--we migrated from `master`->`main` naming. So I was super confused when the agent could not find stuff that I expected to be there.

If it had just asked me first (possibly recommending a default), what branch/ref to create the worktree off of, then the confusion could have been avoided.

---
| Field | Value |
| --- | --- |
| App version | 1.1.3 |
| OS | Windows 10.0.26200 |
| Theme | GitHub |
| Path | /chat |
| Tenure | Day 1 |

まず、暗黙的に worktree を作成するセッションフローを追跡し、その後、複数のリモートと異なるデフォルトブランチを持つリポジトリで再現します。現在、branch/ref の選択がどのように決定されているかを確認します。完了条件は、最初の該当するセッションで worktree のベースブランチまたは ref を尋ね、明確なデフォルトを示し、予期しないブランチを暗黙に選択しないことです。

索引モデルが issue の本文から書いたものです。