What problem does it solve? Drafts shipping under your name often do not sound like you, whether a teammate wrote them, an AI produced them, or you wrote them tired. This Skill reviews a supplied draft against your personal voice profile, counts specific AI-writing tells, rewrites the text in your voice, and separately fact-checks claims against your own captured record without inventing substance. ## Core Features & Use Cases - Tell inventory with evidence tiers: Counts measurable markers such as present participial clauses, nominalizations, style-word verbs, and em dashes, labeling each flag as MEASURED, STRUCTURAL, or CRAFT. - Voice-matched rewrite: Produces a clean version calibrated to your voice skill's register, punctuation fingerprint, and corpus rhythm, with five absolute rules preventing added facts, dropped hedges, or strengthened superlatives. - Fact-check pass: Extracts checkable claims and corroborates them against Littlebird MCP capture, reporting each as Corroborated, Contradicted, Uncorroborated, Unverifiable, or Not checked. - Use Case: A teammate drafts a LinkedIn post going out under your name. Paste it in, and receive a marked-up draft, a counted tell table, a rewrite in your voice, and a list of claims that could not be corroborated against your record. ## Quick Start Paste your draft and ask the assistant to check whether it sounds like you and flag anything it cannot verify against your record.
