Following the recent AI coding discussion, I thought some of you might find this useful, I’ve published the skills i often use focused specifically on AI-assisted audio software development:
Sonic Skills
Precision audio-engineering skills for AI agents
A curated collection of Markdown skills for reviewing, debugging, explaining, and implementing audio software.
The focus is practical:
- Realtime/thread safety
- DSP correctness
- Plugin host compatibility
- Web Audio behavior
- Clicks, dropouts, denormals, unstable filters, broken recall, etc.
Includes skills for:
- DSP + numerics reviews
- JUCE and YUP
- VST3 / CLAP / AU
- Web Audio
- Wwise / FMOD game audio
- DSP algorithm implementation and debugging
Examples:
/audio-dsp-review inspect the current project
/webaudio-review why does this AudioWorklet crackle?
/dsp-algorithm-guide implement a stable SVF lowpass
/clap-review inspect this plugin's process() and params
Designed to compose together, combine DSP safety + numerics + format-specific reviews.
Curious to hear if others are building similar audio-focused workflows around Claude Code, Codex, etc. Also very happy to accept improvements and contributions!