iii-getting-started

This skill is the onboarding guide for iii, a backend engine that positions itself as a single replacement for an API framework, task queue, cron scheduler, pub/sub system, state store, and observability pipeline. It reduces backend development to three primitives — Function, Trigger, and Worker — and the skill walks an AI agent (or developer) through installing the engine, scaffolding a project, starting the engine, and wiring up a first working HTTP endpoint. The problem it solves is the friction of standing up a new backend across disparate services: instead of assembling multiple tools, the user gets one engine and a guided quickstart.

The workflow is a linear six-step path: install the engine via the official install script, run `iii create` to scaffold from a template, start the engine (which listens on a local WebSocket for workers and exposes a REST API and console on localhost ports), install a language SDK (TypeScript, Python, or Rust), register a first function plus an HTTP trigger, and test it with a curl request. It also documents adding pre-built capabilities from a worker registry (`iii worker add`), pinning them in an `iii.lock` file, cron and queue trigger patterns, state persistence via `state::get`/`state::set`, and installing the full set of agent skills. Numerous links point to the upstream iii documentation.

The target users are backend developers and AI coding agents starting a new iii project or needing initial setup and configuration help. It is most useful for first-worker guidance and quickstart scenarios; it explicitly defers to sibling skills (core primitives, SDK reference, engine config, architecture patterns, error handling) for deeper topics, positioning itself as the entry point rather than a comprehensive reference.