fix: force UTF-8 output encoding for PowerShell on Windows

[ysmla]

Issue for this PR

Closes #30205

Related: #30055, #23636

Type of change

• Bug fix

What does this PR do?

解决Windows下，opencode中模型使用powershell输出中文时会出现乱码的问腿

How did you verify your code works?

在中文 Windows 11 + PowerShell 5.1 环境下，用与工具相同的方式启动 PowerShell（管道输出 stdout），会得到 GBK 编码字节（b2e2 cad4 ...），证实了乱码的原因是编码不匹配，修复后，用重新构建的二进制做端到端测试，让模型执行 echo 测试中文：你好世界——打补丁前输出乱码，打补丁后能正确显示中文。
修复前后对比如下：

Checklist

• I have tested my changes locally
• I have not included unrelated changes in this PR

[github-actions]

The following comment was made by an LLM, it may be inaccurate:

Potential Duplicate PRs Found:

1. PR fix: set default UTF-8 encoding for spawned subprocess on Windows #31658 - fix: set default UTF-8 encoding for spawned subprocess on Windows
   fix: set default UTF-8 encoding for spawned subprocess on Windows #31658

   • Directly related: addresses the same Windows encoding issue for subprocesses spawned by opencode

2. PR fix(core): force UTF-8 for Windows shells #36608 - fix(core): force UTF-8 for Windows shells
   fix(core): force UTF-8 for Windows shells #36608

   • Related: appears to be another attempt to fix UTF-8 encoding on Windows shells

These PRs likely cover the same root cause (encoding mismatch when PowerShell outputs on Windows). You should verify if PRs #31658 and #36608 already resolve this issue or if they only partially address it.