DevLLMOps Demo: Quiz App

A quiz app about the DevLLMOps specifications, built following AIgile's methodology + OCPA specs.

Test your knowledge of AI-native software development: team roles, workflow, tooling, and methodology.

Quick Start

cp .env.example .env
make dev
# Open http://localhost:8000

Commands (OCPA)

make dev          # Start dev environment (hot-reload)
make test         # Run integration tests
make prod         # Build and start production
make down         # Stop everything

Project Structure

├── app/
│   ├── Dockerfile         # Multi-stage (dev + prod targets)
│   ├── requirements.txt   # Pinned Python deps
│   └── src/
│       ├── main.py        # FastAPI backend
│       ├── questions.py   # Quiz questions data
│       └── static/        # Frontend (HTML/CSS/JS)
├── scripts/               # OCPA deployment scripts
├── .github/workflows/     # CI: build, test, lint, AI review
├── compose.base.yml       # Shared config
├── compose.dev.yml        # Dev environment
├── compose.prod.yml       # Production environment
├── compose.test.yml       # Integration tests
├── CLAUDE.md              # AI agent context
├── TEAM.md                # Team roster for agent routing
└── Makefile               # Standardized commands

API

Endpoint	Method	Description
/	GET	Quiz frontend
/health	GET	Health check
/api/questions	GET	Questions (without answers)
/api/answers	POST	Submit answers, get score

CI/CD Workflows

Workflow	Trigger	Purpose
Build	Push/PR to main/release	Build dev+prod images, push to ghcr.io
Test	Push/PR to main/release	Integration tests via make test
Linters	Push/PR	Markdownlint, gitleaks, ruff, env validation
AI Review	PR to main	Claude Haiku adversarial code review
PR Direction	PR opened	Block release-to-main PRs

Git Flow (OCPA)

• main -- Staging (auto-deploy)
• release -- Production (manual merge from main)
• Feature branches: short-lived, one per intent/issue
• PR titles prefixed with issue number: #12: Add leaderboard
• Squash merge to main

Methodology

This repo implements the DevLLMOps specification:

• Intent-driven development via GitHub Issues (template)
• AI agent context via CLAUDE.md
• Team routing via TEAM.md
• Automated verification (CI + AI adversarial review)
• Human review only for security-critical and novel changes

License

MIT