Add a systemd user service (spotify-duck.py) that meters Discord's output
stream and the denoised rnnoise mic via pw-record and lowers Spotify's volume
via wpctl while anyone is speaking, restoring it a beat after silence. Tuning
knobs are exposed as service Environment entries. Ignore Python __pycache__.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
