Add config file support for custom voice commands

[turqoisehex]

Summary

hushtype has a hardcoded VOICE_COMMANDS dictionary in hushtype.py (line ~212) that maps spoken phrases to keyboard actions. Users who want custom commands must edit the source code directly. A config file would let users add their own commands without modifying hushtype.py.

The FAQ already mentions this: "Not yet via config file, but the VOICE_COMMANDS dictionary in hushtype.py is straightforward to edit. Custom command support via config file is planned."

Acceptance Criteria

• hushtype reads a JSON or YAML config file (e.g., hushtype_commands.json) on startup
• Config file maps spoken phrases to keyboard actions (e.g., {"bold": "ctrl+b"})
• Built-in commands still work as defaults -- config file overrides/extends them
• If no config file exists, hushtype works exactly as before
• Update README FAQ to reflect the new capability

Relevant Code

• VOICE_COMMANDS dict: hushtype.py line ~212
• _MULTI_STEP dict: hushtype.py line ~257
• _dispatch_command(): hushtype.py line ~271

Notes

This is a good first issue because the command table is already well-structured as a dict. The main work is reading an external file and merging it with the defaults.