Add speech input support

[NullLabTests]

This PR adds speech input support to the Goose CLI using Vosk for offline speech recognition.

Features:

• Toggle voice input mode with /voice command
• Start/stop recording with Ctrl+V
• Uses Vosk for offline speech recognition
• Cross-platform audio capture with cpal
• Seamless integration with existing input system

Testing:

• Tested on Linux
• Voice commands are processed as regular text input
• Works with all existing commands

Dependencies added:

• cpal = "0.15.2"
• whisper-rs = "0.8.0"
• vosk = "0.2.0"

Note: This feature was developed and tested on Linux. Additional platform-specific installation instructions for speech recognition dependencies may be needed.

[michaelneale]

cool! would it be possible to have this on desktop as well? (in some ways perhaps more common to be used there?)

[Kvadratni]

pst https://github.com/Kvadratni/speech-mcp

[NullLabTests]

Nice , converted to draft as it may be redundant?

[baxen]

Wanted to follow up on this! Since this is a bit out of date and we have a few implementations floating around, i'm going to try to consolidate on an issue first and then we can bring things back to PR?

#2952