>_reporules.devOpen Docs
Back to Workflows

Claude Code SaaS Workflow

Production-ready AI coding workflow for Next.js SaaS projects using Claude Code, Cursor and modern repo standards.

Overview

This workflow is designed for AI-assisted SaaS development using:

Next.js App RouterTypeScriptTailwind CSSSupabaseStripeClaude CodeCursor

The goal is to keep AI-generated code predictable, maintainable and consistent across long-term projects.

Recommended Repo Structure

Feature-based architecture prevents AI tools from generating duplicated logic and inconsistent folder structures.

/app
  /(marketing)
  /(dashboard)
  /api

/components
  /ui
  /shared

/features
  /auth
  /billing
  /dashboard
  /settings

/lib
  /db
  /stripe
  /auth
  /utils

/prompts

/rules

/tests

memory.md
rules.md

Architecture Principles

Feature-First Structure

Each business feature owns:

  • components
  • hooks
  • actions
  • schemas
  • tests

Avoid global business logic.

Server-First Architecture

Prefer:

  • Server Components
  • Server Actions
  • async data fetching on server

Avoid unnecessary client-side state.

Shared UI Layer

Reusable UI primitives belong inside /components/ui

Business-specific UI belongs inside /features/*

rules.md

# AI Coding Rules

## Architecture

- Use feature-based architecture
- Keep business logic inside features/
- Shared UI belongs inside /components/ui
- Avoid global utilities unless reusable

---

## React

- Prefer Server Components by default
- Use Client Components only for interactivity
- Never fetch directly inside client components
- Avoid unnecessary useEffect usage

---

## API

- Prefer Server Actions over REST APIs
- Keep route handlers thin
- Centralize validation schemas

---

## Components

- Keep components under 200 lines
- Prefer composition over abstraction
- Avoid deeply nested props

---

## Naming

- Use explicit names
- Avoid generic folders like helpers/, misc/, temp/
- Use feature-oriented naming

---

## Styling

- Use Tailwind utilities
- Reuse shared UI primitives
- Avoid duplicated utility patterns

---

## State Management

- Prefer server state
- Avoid unnecessary global state
- Keep local state isolated

---

## Performance

- Avoid unnecessary client bundles
- Lazy load heavy components
- Prefer streaming when possible

---

## Code Quality

- Avoid duplicated hooks
- Avoid duplicated validation logic
- Keep functions focused and small

memory.md

# Project Memory

## Stack

- Next.js App Router
- TypeScript
- Tailwind CSS
- Supabase
- Stripe

---

## Architecture

- Feature-first structure
- Server Components by default
- Shared UI primitives
- Minimal client-side state

---

## Important Constraints

- Never fetch inside client components
- Never create duplicated hooks
- Never create duplicate folder structures
- Avoid introducing alternative architecture patterns

---

## UI Rules

- Reuse shared UI primitives
- Keep spacing consistent
- Avoid inline styles
- Prefer accessibility-first components

---

## Testing Standards

- Use React Testing Library
- Test user behavior
- Avoid snapshot-heavy tests

---

## Code Style

- Prefer readability over abstractions
- Keep components composable
- Prefer explicit naming

File Organization Standards

Feature Structure

/features/auth
  /components
  /hooks
  /actions
  /schemas
  /tests

Shared UI

/components/ui contains:

buttonsinputsdialogscardstables

Shared Logic

/lib contains:

database clientsauth helpersintegrationsutilities

Prompt Workflow

Generate New Feature

Generate a new dashboard feature following the existing project architecture.

Requirements:

  • Use feature-based folder structure
  • Prefer Server Components
  • Reuse shared UI components
  • Add loading states
  • Add error states
  • Keep components small and composable

Refactor Existing Feature

Refactor this feature without changing business logic.

Requirements:

  • Remove duplicated logic
  • Improve readability
  • Keep server-first architecture
  • Preserve folder structure
  • Reuse existing patterns

Generate Tests

Generate tests following existing testing standards.

Requirements:

  • Use React Testing Library
  • Prefer user behavior testing
  • Avoid snapshots
  • Cover loading states
  • Cover error states

AI Constraints

  • Never create duplicate folder structures
  • Never fetch directly inside client components
  • Never introduce new UI patterns without reuse
  • Avoid unnecessary abstractions
  • Avoid giant utility files
  • Keep feature boundaries strict
  • Prefer explicit naming over generic naming
  • Avoid business logic inside shared components

Testing Workflow

All AI-generated features must include:

  • loading states
  • error states
  • empty states
  • accessibility checks
  • unit tests

Testing Stack

React Testing LibraryVitestPlaywright

Testing Rules

  • Test behavior instead of implementation
  • Avoid excessive mocking
  • Prefer integration-style tests

Pull Request Workflow

Before merging AI-generated code:

  • Check for duplicated logic
  • Check for duplicated hooks
  • Verify folder consistency
  • Verify naming consistency
  • Ensure tests are included
  • Ensure server-first architecture is preserved

PR Guidelines

  • Keep PRs focused
  • Avoid mixed concerns
  • Prefer incremental refactors
  • Keep PRs under 300 LOC when possible

Before Applying Workflow Standards

  • duplicated hooks
  • mixed server/client logic
  • inconsistent folder structures
  • large components
  • duplicated utility files
  • scattered business logic
  • predictable architecture
  • reusable UI patterns
  • cleaner AI-generated code
  • smaller components
  • consistent folder structure
  • easier long-term maintenance

Recommended Stack

Frontend

Next.js App RouterTailwind CSSshadcn/ui

Backend

SupabasePostgreSQL

Payments

Stripe

Deployment

Vercel

AI Tools

Claude CodeCursor

Workflow Goals

This workflow is designed to:

  • reduce inconsistent AI output
  • improve long-term maintainability
  • keep repo structure predictable
  • improve AI-assisted collaboration
  • reduce duplicated logic
  • scale AI coding across teams

Copy This Workflow

Save these files to your repo and reference them in your AI tool configuration.