What problem does it solve? Writing READMEs that stay accurate and useful is hard: sections go stale, setup instructions get duplicated across a monorepo, and different project types need different structures. This Skill provides a decision framework, per-project-type templates, and anti-pattern rules so every README matches its audience. ## Core Features & Use Cases - Project-type detection and templates: Identifies whether the target is a T3 monorepo app, internal package, CLI tool, OSS library, personal project, config directory, or internal service, then loads exactly one matching template. - Judgment framework and decision matrix: Applies audience, staleness-risk, and single-source tests plus a section matrix to decide what belongs in each README. - Reference depth on demand: Loads Standard README spec, examples, and README philosophy references only when the task requires compliance or deeper guidance. - Use Case: Ask for a README for a new package in packages/ and receive a scoped document with exports table, usage example, and gotchas, without duplicated root-level setup instructions. ## Quick Start Ask the agent to write a README for your project, for example: write a README for the packages/db workspace package in my T3 monorepo.
