Spamming server errors during subagent execution

### Describe the bug

There are a lot of transient server errors, especially during subagent tasks. Every time I run anything using fleet or anything that requires a subagent task, I'll get 10-15 or more copies of "Response was interrupted due to a server error. Retrying...".

Nothing seems to be affected, and the /tasks read-only view does not show any issues.

### Affected version

GitHub Copilot CLI 1.0.74-0

### Steps to reproduce the behavior

_No response_

### Expected behavior

It would be nice if the messages could be collapsed into one, for example, that updates a disruption count in place, or even just a little "Unstable connection, 5 retries" at the bottom somewhere.

### Additional context

_No response_

首先在 GitHub Copilot CLI 1.0.74-0 上重現一個 fleet 或 subagent 任務，並搭配唯讀的 /tasks 檢視畫面觀察重試訊息。追蹤重複的 "Response was interrupted due to a server error. Retrying..." 訊息在哪裡發出；當暫時性重試以一則摺疊訊息或中斷計數表示，且不改變任務行為時，即完成。

由索引模型根據 Issue 內容生成。