What is vibe-coding?

Vibe-Coding is not a piece of software you install. It is a written guide that lays out best practices for coding with the help of AI agents, a style often called vibe coding where you describe what you want in plain English and let an AI write, edit, and debug code for you. The guide's core idea is that AI should never make the final call on its own. Instead the recommended workflow is: AI generates code, a human reviews it, the AI improves it based on feedback, and only then does a human deploy it. The author argues that blindly trusting AI generated code without review tends to introduce security and quality problems. Most of the document is a set of practical habits. It suggests giving the AI full context about your project structure, coding standards, and existing files rather than vague one line requests. It recommends breaking large tasks into small steps instead of asking for an entire app at once, editing specific files rather than regenerating whole projects, committing changes often so mistakes can be undone, and running a security scan after every AI generated change to catch things like exposed secrets or unsafe file uploads. The guide also includes checklists covering security, such as password hashing and rate limiting, performance, such as caching and lazy loading, and interface design, such as responsive layouts and accessibility. It lists common mistakes to avoid, like skipping tests or hardcoding API keys, and closes with a table of popular tools people commonly use when vibe coding, including various AI models, code editors, backend frameworks, and deployment platforms. Overall this repository works like a checklist and mindset primer for anyone using AI coding assistants, aimed at helping them ship more reliable software instead of shipping whatever the AI produces first.