The Prompt Framework: How to Write Prompts That Work
A vague prompt leaves Claude Code to decide the details, and its defaults aren't yours. The framework keeps those decisions with you: what you want, where it goes, and how it should behave.
The problem with vague prompts
You type a request, and Claude Code builds something that answers it but isn't quite what you meant. You clarify. It adjusts. You clarify again. Twenty minutes later you have something usable, and most of that time went into wrangling Claude Code. The trouble started with your first message.
Ask a person to 'add a button,' and they'll come back with questions: Where? What color? What does it do? Someone who's worked with you for a while fills in those gaps on their own, because they know your project, your preferences, and the tools you use.
Claude Code doesn't have that background. Each conversation starts fresh. Write 'add a button to the form,' and it still has to settle which form, what the label says, which building blocks to use, where in the file it goes, and what happens on click.
That's the vague prompt trap. You can settle those details once, in your prompt, or correct them one round at a time afterward.