# Setup check — run this before the workshop

A preflight "doctor" that verifies your machine is ready for **Agentic AI Architect**.
It checks the tools the labs actually need, prints clear ✓ / ⚠ / ✗ results, and (opt-in)
installs the ones a package manager can handle.

## Run it

**macOS / Linux**
```bash
chmod +x setup-check.sh
./setup-check.sh             # check + guidance
./setup-check.sh --install   # also install what it can (Homebrew / apt / dnf, + uv)
./setup-check.sh --strict    # treat warnings as failures too (for facilitators)
```

**Windows (PowerShell)**
```powershell
.\setup-check.ps1            # check + guidance
.\setup-check.ps1 -Install   # also install what it can (winget, + uv)
.\setup-check.ps1 -Strict    # treat warnings as failures too (for facilitators)
# if scripts are blocked:
powershell -ExecutionPolicy Bypass -File .\setup-check.ps1
```

**Exit codes:** `0` = core toolchain present (warnings allowed); `1` = a required tool is
missing — or, with `--strict` / `-Strict`, any warning. Participants want green-or-warnings;
facilitators prepping a lab image can use strict to require a fully clean run.

## What it checks
- **Node.js v20+**, **npm**, **npx** — required (the MCP server + the Day-2 app)
- **git** + a GitHub identity — you push your repo as the portfolio artifact
- **uv / uvx** — for the DuckDuckGo search server (Lab 1). *Not blocking* — the offline
  `mock-search` server is the fallback.
- **Antigravity** (`agy` on PATH) — the workshop host. **Manual install** (can't be scripted).
- **npm registry reachable** + **ports 3000 / 3001 free** (the Day-2 frontend/backend).

## What it can NOT do for you
- Install **Antigravity** (App + CLI) — get it from the workshop's delivery instructions.
- Create **accounts** (GitHub, and any cloud accounts your track uses).
- Pull **Ollama** models (optional offline fallback).

A green run means the toolchain is ready; finish the manual items and you're set.