What problem does it solve? Agents drafting prose on your behalf sound generic because they lack a grounded model of how you actually write. This Skill builds a reusable voice-profile skill from your real writing so agents can draft and edit text that sounds like you. ## Core Features & Use Cases - Multi-source corpus gathering: Collects your writing from GitHub PRs and comments, Google Docs, local files, session transcripts, Slack, and email, with rate-limit-safe fetching via bundled scripts. - Stylometric analysis with AI-content filtering: Computes em-dash, semicolon, sentence-length, and contraction metrics per register while excluding agent-generated content that would poison the profile. - Voice profile synthesis: Produces a personalized SKILL.md with verbatim examples, register playbooks, signature phrases, anti-patterns, and a pre-send checklist. - Use Case: Ask an agent to learn your style from your GitHub PRs and design docs; it fetches your authored content, filters out AI-written passages, and outputs a sound-like-me skill any agent can load to write in your voice. ## Quick Start Ask the agent to build a write-like-me skill from your GitHub login and orgs, your blog folder, and your session transcripts.