jchook/shhh — reverse-engineered prompt

Reverse engineered prompt

Build me a small command line app called shhh that listens to my computer microphone and lets me know when I am being too loud. I want a simple flow where I can run a calibrate command first so it learns my room and mic, then start the app and have it watch the input level and play a “SHHH!” sound whenever I go over a decibel threshold.

Please also add a way to list available input devices, choose a specific mic, and generate a starter config file with comments. Settings should be usable from command line options, environment variables, or a config file, with command line taking priority. I want sensible controls for alert frequency, loudness threshold, sensitivity to spikes versus sustained noise, system notifications, verbosity, and a custom alert sound file.

If possible, make it work on Linux, Mac, and Windows, and keep the setup easy enough that I can just install it and run it. You can look up current audio and notification docs online if needed.

Want more depth? Deep Reverse