new-project
Scaffold a new full-stack project with backend (Scalingo) and Next.js frontend (Cloudflare Pages), initialize git/GitHub, and plan using GitHub issues.
Best use case
new-project is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Scaffold a new full-stack project with backend (Scalingo) and Next.js frontend (Cloudflare Pages), initialize git/GitHub, and plan using GitHub issues.
Teams using new-project should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/new-project/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How new-project Compares
| Feature / Agent | new-project | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Scaffold a new full-stack project with backend (Scalingo) and Next.js frontend (Cloudflare Pages), initialize git/GitHub, and plan using GitHub issues.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# New Project Scaffold and bootstrap a new full-stack project. This skill is both the project creator and the foundation for how the project is planned and tracked — using GitHub issues as the planning database. ## Input Handling - **Argument required**: project name (kebab-case), e.g. `/new-project my-app` - Optionally followed by a description: `/new-project my-app "A tool that does X"` - If no argument, ask the user for project name and description ## Your Task ### 1. Gather Information Ask the user using AskUserQuestion: - **Project description** (if not provided as argument) - **Backend stack**: Python (FastAPI) or Node.js (NestJS) - **Scalingo region**: osc-fr1 (default) or osc-secnum-fr1 - **GitHub visibility**: Public or private ### 2. Create Project Structure ``` ~/dev/<project-name>/ ├── api/ # Backend ├── web/ # Next.js frontend ├── docs/ # Product documentation (empty) ├── CLAUDE.md # Project instructions for Claude Code ├── PLAN.md # Agent team workflow └── README.md ``` ### 3. Backend Setup (api/) #### If Python (FastAPI): Scaffold a minimal FastAPI app with: - `uv` as package manager (pyproject.toml, uv.lock) - Python 3.12 (.python-version) - `Procfile` for Scalingo: `web: PYTHONPATH=. python -m uvicorn api.main:app --host 0.0.0.0 --port $PORT` - Minimal FastAPI app entry point with a `/health` endpoint - Dev tooling: `ruff` (linter/formatter), `ty` (type checker), `pytest` - Standard .gitignore for Python #### If Node.js (NestJS): Scaffold a minimal NestJS app with: - `pnpm` as package manager - `Procfile` for Scalingo: `web: node dist/main.js` - Minimal NestJS app with a `/health` endpoint - Dev tooling: `eslint`, `prettier` - Standard .gitignore for Node ### 4. Frontend Setup (web/) Scaffold a Next.js 16 app with: - `pnpm` as package manager - TypeScript, Tailwind CSS 4, shadcn/ui - OpenNext.js + Cloudflare Workers adapter (wrangler.toml, open-next.config.ts) - Dev tooling: `eslint`, `prettier` - Scripts: `dev`, `build`, `preview` (opennextjs-cloudflare), `deploy` (opennextjs-cloudflare) - `.env.local` with `NEXT_PUBLIC_API_URL=http://localhost:8000/api` - `.env.production` with `NEXT_PUBLIC_API_URL=https://<project-name>-api.<region>.scalingo.io/api` - Minimal API client in `src/lib/api.ts` using `NEXT_PUBLIC_API_URL` - Standard .gitignore for Next.js + Cloudflare ### 5. Root Files - **CLAUDE.md**: Project-level instructions including: - Project overview and tech stack - How to run backend and frontend locally - Deployment targets (Scalingo for api/, Cloudflare Pages for web/) - Lint/test commands for each stack - **README.md**: Project title, description, tech stack table, quick start commands - **PLAN.md**: Agent team workflow with 4 phases: 1. Discovery (product + devils-advocate) 2. Design (product + devils-advocate) 3. Build (backend-dev + frontend-dev, supervised by devils-advocate) 4. Validate (devils-advocate) Adapt roles and deliverables to the specific project. ### 6. Initialize Git and GitHub 1. `git init` 2. `git add -A && git commit -m "feat: scaffold project structure"` 3. `gh repo create <project-name> --<visibility> --source . --push` ### 7. Plan Using GitHub Issues Use GitHub issues as the planning database. Create an initial set of issues based on the PLAN.md phases using `gh issue create`. The agent should document its plan, decisions, and progress as GitHub issues — not just at project creation but as an ongoing practice. Issues should be created and updated as the project evolves. ### 8. Install Dependencies Run in parallel: - Backend: `uv sync` (Python) or `pnpm install` (Node) - Frontend: `pnpm install` ### 9. Final Report Print a summary with project location, GitHub repo URL, created issue URLs, and quick start commands. ## Guidelines **DO:** - Keep the scaffold minimal - just enough to run and deploy - Use modern tooling: uv/ruff/ty (Python), pnpm/eslint/prettier (Node/TS) - Create a CLAUDE.md so Claude Code understands the project from the start - Use GitHub issues as the single source of truth for planning - Create the `docs/` directory (empty, ready for specs) **DON'T:** - Don't add business logic - just the bare scaffold - Don't deploy anything - just set up the structure ## Example Usage ``` /new-project my-app /new-project my-app "A marketplace for vintage furniture" ```
Related Skills
ln-700-project-bootstrap
Orchestrates full project bootstrap from Replit export to production-ready structure
init-project
当用户明确要求"初始化项目"、"创建项目指令文件"或"生成 AGENTS.md"时使用。完全自动化:自动检测操作系统默认语言,分析项目目录结构(支持 Python/Web/Rust/Go/Java/数据科学/文档项目等),推断项目类型和用途,一键生成规范的项目指令文档。生成文件包括:AGENTS.md(跨平台通用项目指令,Single Source of Truth)、CLAUDE.md(Claude Code 特定适配,通过 @./AGENTS.md 引用)、README.md(项目介绍与使用方法)、CHANGELOG.md(项目变更记录)。
go-project-overview
Overview of Go project structure, technology stack (Echo, Connect RPC, Driver pattern), and essential commands for building, testing, and development
fastapi-project
Scaffold and evolve FastAPI projects with uv-based tooling, structured settings, and production-ready observability, resilience, availability, and security patterns aligned with python.instructions.md.
custom-project-standards
Hệ thống tiêu chuẩn dự án đa năng (Standard Platform). Hỗ trợ Frontend, Backend, DevOps với nhiều tùy chọn ngôn ngữ/framework.
create-spring-boot-kotlin-project
Create Spring Boot Kotlin Project Skeleton
create-spring-boot-java-project
Create Spring Boot Java Project Skeleton
chatgpt-project-architect
Use cuando necesites crear o mejorar instrucciones de proyecto ChatGPT, diseñar system prompts, definir agentes conversacionales, estructurar modos y comandos, implementar anti-injection, o validar calidad de project instructions. Keywords: chatgpt project, system prompt, project instructions, agent design, prompt engineering, anti-injection, command shortcuts, agent modes.
cc-skill-project-guidelines-example
Project Guidelines Skill (Example)
awesome-copilot-root-create-spring-boot-java-project
Create Spring Boot Java Project Skeleton Use when: the task directly matches create spring boot java project responsibilities within plugin awesome-copilot-root. Do not use when: a more specific framework or task-focused skill is clearly a better match.
android-project
Navigate and analyze Android project structure, modules, and dependencies. Use when exploring project structure, finding related files, analyzing dependencies, or locating code patterns.
analyzing-projects
Analyzes codebases to understand structure, tech stack, patterns, and conventions. Use when onboarding to a new project, exploring unfamiliar code, or when asked "how does this work?" or "what's the architecture?"