Project Tools & Guides

40 guides covering tech stacks, kickoff prompts, best practices, and essential files

Showing 40 guides

Tech StacksMost Popular

The "Vibe Coder" Stack

Most popular stack for AI-assisted projects. Zero backend code needed with generous free tiers.

Tech StacksAdvanced

The "Full Control" Stack

Full control over your backend for complex business logic and custom APIs.

Tech StacksFastest

The "Rapid Prototype" Stack

Go from idea to deployed app in hours. Perfect for MVPs and hackathons.

Tech StacksMobile

The "Mobile App" Stack

One codebase for iOS and Android with great developer experience.

Tech Stacks

Project Folder Structure Template

Recommended folder structure for AI-assisted projects with all essential files.

Kickoff PromptsEssential

"Day Zero" Project Onboarding Prompt

Paste into Claude Code or Cursor when starting a brand new project. Gets you a complete scaffold in minutes.

Kickoff PromptsComplete Plan

Project Kickoff Master Prompt (13 Deliverables)

Comprehensive planning prompt that produces a complete project plan before you write any code.

Best Practices

Start with a Clear Vision Document

Before any code, write down what your app does, who it serves, and core features. Share with AI as the first message.

Best Practices

Design Your UI/UX Before Coding

Sketch screens on paper, Figma, or Excalidraw. AI generates much better code with a visual target.

Best Practices

Set Up Git from Minute One

Initialize a repo immediately and commit after every working change. This is your safety net.

Best Practices

Lock In Your Tech Stack Early

Choose once and stick with it. State your stack in CLAUDE.md so AI never suggests alternatives.

Best Practices

Write Detailed AI Rules Files

Your CLAUDE.md or .cursorrules is the most important file in your project. Update it as you evolve.

Best Practices

One Feature, One Prompt

Never ask AI to "build the whole app." Break features into the smallest possible units.

Best Practices

Always Provide Context in Prompts

Start each prompt with what file you're in, what you've built, and what the expected behavior is.

Best Practices

Review Every Line of AI Code

Never blindly accept AI output. Read through it and understand what it does. You are the quality gate.

Best Practices

Test After Every Change

After each AI-generated change, run the app and test the specific feature. Don't stack untested changes.

Best Practices

Fix Bugs Immediately

One unfixed bug leads to AI building more code on top of broken foundations. Fix it now.

Best Practices

Use Plan Mode for Complex Features

For anything touching multiple files, ask AI to plan first. Review, adjust, then implement step by step.

Best Practices

Keep Components Small and Focused

If a file grows past 150-200 lines, split it. Small components are easier for both humans and AI to work with.

Best Practices

Never Commit Secrets or .env Files

Use .env.local for development and .env.example for templates. AI sometimes suggests hardcoding secrets.

Best Practices

Build Mobile-First from the Start

Start with mobile layout, then add responsive breakpoints. Retrofitting mobile is painful.

Best Practices

Use AI for Debugging Too

When stuck, paste the error and code to AI. Describe what you expected vs what happened.

Best Practices

Don't Fight the Framework

Use Next.js patterns with Next.js. Don't ask AI to implement patterns from other frameworks.

Best Practices

Document As You Go

Add comments to complex logic and keep AI context files updated. Future-you will thank present-you.

Best Practices

Ship Early, Iterate Often

Deploy as soon as you have a working MVP. Real user feedback beats perfecting features in isolation.

Project Files

CLAUDE.md

The context file for Claude Code. Automatically read at the start of every session, giving Claude persistent project memory.

Project Files

AGENTS.md

Defines how AI agents should behave — permissions, workflow rules, and specialized instructions.

Project Files

skills.md

Reusable prompt patterns and common operations that AI assistants can draw from.

Project Files

CHANGELOG.md

Chronological log of notable changes, organized by version or date.

Project Files

FEATURES.md

Living document listing all planned, in-progress, and completed features.

Project Files

PROJECT.md

High-level architectural document explaining the "why" behind your design decisions.

Project Files

README.md

The front door of your project. The first file anyone (or any AI) reads.

Common Pitfalls

Don't Ask AI to "Build the Whole App"

Break it into features and implement them one at a time. Massive prompts produce massive bugs.

Common Pitfalls

Don't Skip Understanding the Code

If you can't explain it, you can't debug it. Take time to understand what AI generates.

Common Pitfalls

Don't Ignore Errors or Warnings

Fix them immediately. Technical debt compounds fast with AI-generated code.

Common Pitfalls

Don't Commit Secrets or .env Files

Never commit API keys or passwords. Use .env.local and .gitignore.

Common Pitfalls

Don't Skip Mobile Responsiveness

Test on mobile from the start. Retrofitting responsive design later is painful.

Common Pitfalls

Don't Let AI Switch Your Tech Stack

AI sometimes suggests different libraries mid-project. Lock your stack in context files.

Common Pitfalls

Don't Stack Untested Changes

Test after every AI-generated change. When something breaks, know exactly which change caused it.

Common Pitfalls

Don't Forget to Update AI Context Files

Your CLAUDE.md or .cursorrules should evolve with your project. Stale context = stale output.

Have a project tip to share?

Submit your favorite tech stacks, kickoff prompts, or best practices

Submit a Guide

Command Palette

Search for a command to run...