An interactive fleet and service-queue management web application built with Angular 21 (Signals, resource, linkedSignal, Angular Aria, SSR with Express) and Genkit (genkit + @genkit-ai/google-genai powered by gemini-3-flash-preview).
This repository accompanies the live stage demo showing how to pair-program with an AI coding agent using Agent Skills, MCP Servers, and Antigravity Plugins (chrome-devtools-plugin).
You can either start on main and run each prompt in content/prompts/ with your coding agent, or check out any of the completed checkpoint branches directly:
| Branch | Prompt | Description |
|---|---|---|
main |
— | Clean starter baseline with fleet map, inventory list, service queue, and static chat shell. |
step-1 |
content/prompts/prompt-1.md |
Initializes Genkit with googleAI() (gemini-3-flash-preview) and adds the POST /api/chat endpoint in src/server.ts. |
step-2 |
content/prompts/prompt-2.md |
Connects the client chat UI (src/app/components/chat/chat.ts) to /api/chat using the Angular resource API and afterRenderEffect auto-scrolling. |
step-3 |
content/prompts/prompt-3.md |
Defines 7 Genkit fleet & service-queue tools (ai.defineTool) in src/server.ts and synchronizes UI state using Angular resource, linkedSignal, and .reload(). |
To jump straight to the completed application with full Genkit tool calling:
git checkout step-3npm installCreate a .env file in the project root:
GEMINI_API_KEY="your_gemini_api_key_here"If you want to run the prompts in content/prompts/ with an AI coding agent from main:
-
Install the Angular & Genkit Agent Skills:
npx skills add https://github.com/angular/skills --skill angular-developer npx skills add firebase/genkit --skill developing-genkit-js
-
Install the Chrome DevTools Plugin (in Antigravity CLI):
/plugin install chrome-devtools-plugin -
MCP Configuration:
- Project MCP settings are included in
.gemini/settings.json, and agent workspace guidelines are defined inAGENTS.md.
- Project MCP settings are included in
To start the local development server:
npm startOpen http://localhost:4200/ in your browser.
To build the project for production:
npm run build