`gh stack view` silently ignores positional arguments, including a valid different stack number

从 gh stack view 命令的参数解析开始，复现 issue 中的两个命令，包括它们的退出状态和输出。检查 gh stack checkout 如何处理 stack 编号，然后为有效和无效的位置参数添加覆盖测试。完成的标准是该命令不再静默忽略这些参数：它应该明确拒绝这些参数，或一致地支持 stack 选择。

Summary

gh stack view accepts no positional argument, but supplying one is silently discarded instead of rejected. The harmful case is passing a valid stack number that isn't the current one — you get a different stack's contents, with that stack's number in the header, and no indication your argument was ignored.

On a branch belonging to stack #6, asking for stack #3 (which exists):

$ gh stack view --short 3
Stack #6                          <-- rendered #6, not the #3 that was requested
├ stack-b-two ○ #5
» stack-b-one ○ #4 (current)
└ main

$ gh stack view --short garbage999
Stack #6                          <-- also silently succeeds
├ stack-b-two ○ #5
» stack-b-one ○ #4 (current)
└ main

Exit status is 0 in both cases.

Why this matters

The output carries a stack number in its header, so it reads as authoritative. Anyone comparing two stacks — or scripting against view — can act on the wrong one without any signal. It is also a natural mistake to make, because gh stack checkout does take a stack number, so the two commands look like they share a selector.

Reproduction

In a repo with two stacks:

git switch <a-branch-in-stack-B>
gh stack view --short <stack-A-number>   # shows stack B
gh stack view --short garbage999         # shows stack B

Public repo with two stacks set up for this: https://github.com/xn/gh-stack-repro

Expected

Either error with unknown argument "3" (the usual gh behaviour for unexpected positionals), or accept a stack selector here the way gh stack checkout does. The latter would also give a non-interactive escape hatch for the ambiguity in #415.

Environment

• gh 2.97.0
• gh-stack v0.1.0
• git 2.50.1
• macOS 26.4.1, darwin/arm64