opencode-plugin-encoding-auto

OpenCode plugin that auto-detects file encoding (EUC-KR, CP949, Shift-JIS, GB2312, etc) and handles read/edit/write without garbled text on Windows.

Problem

On Windows, when working with files saved in non-UTF-8 encodings (e.g. Korean EUC-KR, Japanese Shift-JIS, Chinese GB2312), opencode's built-in tools produce garbled text (mojibake) because they assume UTF-8.

Solution

This plugin:

1. Read — Auto-detects file encoding using chardet, decodes with iconv-lite, and outputs proper text to the console.
2. Edit/Write — Temporarily converts the file to UTF-8 so opencode's built-in edit/write tools run normally (including diff display), then converts back to the original encoding after the operation completes.
3. Bash (PowerShell) — Prepends [Console]::OutputEncoding = [System.Text.Encoding]::UTF8 to PowerShell commands for proper console output.

Installation

Add to your opencode.json:

{
  "plugin": ["opencode-plugin-encoding-auto"]
}

Or use a local path:

{
  "plugin": ["./plugins/encoding-auto.ts"]
}

Dependencies

• chardet — character encoding detection
• iconv-lite — encoding/decoding

How It Works

tool.execute.before (edit/write):
  file.enc  →  detect encoding  →  convert to UTF-8 temp
                                         ↓
                              built-in edit/write runs (diff shown)
                                         ↓
tool.execute.after (edit/write):
  file.utf8  →  read UTF-8  →  convert back to original encoding  →  save

tool.execute.after (read):
  file.enc  →  detect encoding  →  decode  →  output to console

Supported Encodings

Any encoding supported by iconv-lite, including:

• EUC-KR / CP949 (Korean)
• Shift-JIS / CP932 (Japanese)
• GB2312 / GBK (Chinese)
• ISO-8859-1 / Windows-1252 (Western)
• And many more

License

MIT