# Shoplazza AI Toolkit overview

> Shoplazza AI Toolkit lets AI coding agents run a store in plain language, pairing Shoplazza CLI as the execution engine with per-domain Agent Skills.

The Shoplazza AI Toolkit lets AI coding agents operate the Shoplazza platform in natural language. It is built on [Shoplazza CLI](/docs/app/developer-tools/shoplazza-cli/), which is **agent-native**: every command returns structured JSON out of the box, so an agent can manage products, discounts, orders, and customers with zero extra setup.

## How it works

The toolkit has two parts:

1. **Shoplazza CLI** — the execution engine. It authenticates against a store and exposes a [three-layer command system](/docs/app/developer-tools/shoplazza-cli/guides/command-layers) (shortcuts → API commands → raw API) with structured JSON output.
2. **Agent Skills** — `SKILL.md` guidance files, one per domain, that teach your agent *when* and *how* to call each command. You install them with a single command and invoke them with a `/shoplazza-<domain>` prefix.

Once both are in place, you describe a task in plain language and your agent runs the right command for you:

```
/shoplazza-products Add a new product: Summer print T-shirt, priced 129
```

## What you can do

## Next steps

* [Get started](./get-started) — Install, add skills, authenticate, and run your first prompt
* [Agent Skills](./skills) — How skills, prefixes, and the safety protocol work
* [Safety & permissions](./safety-and-permissions) — Scopes, the safety protocol, and guardrails
* [Usage examples](./usage/products) — Prompt library by module (products, discounts, orders, customers, shop, webhook, billing)
* [End-to-end workflow](./end-to-end) — A full campaign across multiple domains
* [Headless & CI](./headless-ci) — Run the toolkit in automated environments
* [Troubleshooting](./troubleshooting) — Fix common agent and auth issues