genideva/voice-profile
Measure your writing voice from your own pre-AI writing and calls, score your Big Five with quoted evidence, and get a prompt plus a pass/fail verifier that holds AI to your voice.
Build a measured writing-voice profile from the user's own pre-AI writing and call transcripts, then turn it into a system prompt that makes AI write in their voice. Use when someone wants AI to "sound like me", wants their voice analyzed or measured, wants Big Five or personality scores from their writing, or wants a voice/style prompt built from samples. Walks through corpus intake, register detection, surface metrics, Big Five scoring with quoted evidence, shaping prompt, and a verify loop.
Check whether a draft sounds like the user, using the voice profile built with the voice-profile build skill. Use when the user asks "does this sound like me", wants a draft scored against their voice, wants an AI draft rewritten into their voice, or wants to write something new in their voice. Requires profile.json and shaping-<register>.md from a previous build.