An AI agent skill that scaffolds new Django projects following a consistent set of production-ready patterns.
npx skills add dr-rompecabezas/django-scaffold-skillThis works with Claude Code, Codex, Cursor, and 40+ other agents.
Then invoke it inside your agent session:
/django-scaffold
Given a directory with uv init done and Django installed, this skill asks 7 targeted questions and generates a complete, production-ready project structure — every file, configured correctly, with no placeholders.
- You are already inside the project directory
uv inithas been run —pyproject.tomlexists.venvis created and activateddjango(ordjango+wagtail) is already installed viauv add
| # | Question | Options |
|---|---|---|
| Q1 | Project type | Django full-stack / Wagtail CMS / Django REST API |
| Q2 | Project name | Used for pyproject.toml, Railway service, and DB name |
| Q3 | Main app name | Primary local app beyond users (default: core) |
| Q4 | Celery + Redis? | yes / no |
| Q5 | Frontend interactivity? | HTMX / Alpine.js / both / neither (full-stack only) |
| Q6 | Authentication | django-allauth / allauth + Google OAuth / built-in auth |
| Q7 | Languages? | English only or multilingual list |
Root files: manage.py, pyproject.toml (updated), .env, .env.example, .gitignore, .pre-commit-config.yaml, railpack.json, railway.json, docker-compose.yml, package.json (full-stack only)
config/: settings/base.py, settings/local.py, settings/production.py, urls.py, wsgi.py, asgi.py, celery.py (if Q4=yes)
users/: Custom User model, admin, migrations, test factories
{app_name}/: Models, views, URLs, admin, migrations, tests scaffold
templates/: base.html with Tailwind CSS, HTMX, and Alpine.js wiring (full-stack only)
.github/workflows/: ci.yml with pre-commit + pytest
| Area | Decision |
|---|---|
| Deployment | Railway |
| Media storage | S3 with presigned URLs (django-storages[s3]) |
| Error monitoring | Sentry SDK |
| CI/CD | GitHub Actions |
| Styling | Tailwind CSS v4 via npm (no config file needed) |
| Python | latest stable |
| Django | latest LTS (x.2 series) |
| Dependency management | pyproject.toml + uv |
| Local services | Docker Compose (Postgres, pgAdmin, Mailpit, Redis/Celery if needed) |
| Django itself | Runs directly in terminal — not containerized |
| File | Purpose |
|---|---|
skills/django-scaffold/SKILL.md |
Skill definition — question flow and generation checklist |
skills/django-scaffold/reference.md |
Authoritative templates and patterns for all generated files |
This project is licensed under the MIT License.
The interview questions style is inspired by Cookiecutter Django, but with an AI-driven twist (directive but not deterministic) and far narrower focus.
Much of the developer experience tooling (e.g., pre-commit) is inspired by Adam Johnson's books and blog posts on production Django patterns.