mikeryandev/miniature-octo-couscous

This repository is a hands-on workshop for frontend engineers who want to practice using AI coding tools on a realistic project. The task is to build the interface for a shared to-do list app where people can log in, create task lists, invite collaborators, assign to-dos to each other, check items off, and see live updates as things change. The repo gives you a working backend with a fully specced-out API, but leaves the frontend almost entirely to you. The backend handles authentication, data storage, and real-time updates. It runs in memory, so restarting it wipes everything clean. The API speaks plain HTTP for actions like creating lists and completing tasks, plus WebSockets for live event streaming, so whenever someone adds or finishes a to-do, everyone connected sees it instantly. The session cookie from login also authenticates the WebSocket connection, keeping things simple on the security side. Error responses follow a consistent format with machine-readable codes, making them straightforward to handle in code. The target user is an experienced frontend developer who already knows TypeScript and modern tooling but wants to get better at working alongside AI coding agents. Instead of a toy problem, they get a real product surface with auth, permissions, real-time sync, and validation errors to handle. The workshop is designed to let them focus on app architecture, state management, API integration, accessibility, and testing rather than spending time scaffolding a backend. A notable design choice is that the starter frontend intentionally lacks an API client and dev proxy. Participants are expected to wire up same-origin requests themselves, which mirrors the kind of decisions you face in actual projects. The permission model is also well thought out: list creators are automatically owners who can manage members, members can create and assign tasks, and non-members simply see nothing, with the API returning generic "not found" responses to avoid leaking information about private lists.