aaddrick-voice-replication

Invoke this Skill when asking the AI to draft content, and it applies codified writing rules, few-shot examples, and numeric style targets derived from a corpus analysis. The output mimics sentence rhythm, tone, and structure rather than just topic content.

The Skill validates output against measurable targets like Flesch-Kincaid grade 6-8, 8-12 word average sentence length, ~2% contraction rate, and sentiment compound of +0.15 to +0.40. A self-verification checklist runs before finalizing output.

A voice agent is a persona the AI adopts to write as the person, while a voice skill is a rule set the AI executes to replicate the voice. This Skill provides the rules, examples, and numeric profile; the companion agent can reference it for constraints.

Avoid it for formal, academic, legal, or corporate writing, since this voice is conversational and calibrated for informal technical and personal contexts. Also skip it when the user explicitly requests a different tone or style.

Yes, the Skill defines three registers: technical (advisory tone, 10-14 word sentences), personal/parenting (warmer, longer narrative sentences), and hobby/casual (shortest sentences, most contractions). It defaults to technical-advisory when context is ambiguous.