readme-structurelisted

## Purpose A README is usually the first (and sometimes only) doc a new contributor or user reads. Use this skill when creating a new README from scratch, restructuring one that's grown unwieldy, or deciding whether new content belongs in the README or needs its own page. ## When to use it - Starting a new project or package that doesn't have a README yet. - A README has become long enough that readers are visibly skipping sections or asking questions the README already answers (a sign of structure, not content, failure). - Deciding where a new piece of documentation belongs — extend the README, or start a docs/ page. ## Default README shape 1. **What it is** — one or two sentences: what the project does and who it's for. No setup instructions yet, no history, just the summary a stranger needs before deciding to keep reading. 2. **Quickstart** — the shortest path from "nothing installed" to "it's running." Concrete, copy-pasteable commands, not a description of the steps. If there are prerequisites, list them right before the commands that need them, not in a separate section the reader has to hunt for. 3. **Key commands** — the small set of commands a working contributor actually uses day to day (run tests, run the linter, start the dev server, build). Not an exhaustive CLI reference — link out to that if it exists. 4. **Where to find more** — links to deeper docs: architecture notes, contribution guidelines, the agent-facing instruction file if there is one, API refere