What problem does it solve? Capturing how a person, brand, or character actually writes is hard because self-descriptions are aspirational and unreliable. This Skill analyzes a real writing corpus, validates hypotheses through a structured gap interview, and compiles a portable voice profile under 5,000 tokens that an AI can use as standing context. ## Core Features & Use Cases - Corpus-Based Extraction: Analyzes 12 behavioral dimensions (sentence rhythm, pronoun stance, punctuation habits, humor register, and more) to build 8-12 falsifiable, evidence-backed hypotheses instead of asking how the subject sounds. - Type-Specific Gap Interviews: Runs distinct interview tracks for Person, Brand, or Role/Character subjects, covering refusals, red flags, context switches, and decision rules that text alone cannot reveal. - Compression and Validation: Compiles findings into a structured XML voice profile with a source hierarchy (corpus-observed beats self-described), then validates it with a side-by-side voice test before delivery. - Use Case: A founder wants AI-generated content to sound like them. Provide 5 past emails and essays; the Skill deduces their voice patterns, confirms them one at a time, and outputs an about-me.md profile ready to use as context for any writing task. ## Quick Start Analyze these five emails I have written and compile a personal voice profile that captures how I actually write.
