Writing Your First Effective Prompt

Prompting Claude Code vs Chat AI

Claude Code prompts are more like instructions to a senior engineer than chat messages. You don't need to explain what files to read — Claude will find them. Focus on the outcome, not the steps.

The Anatomy of a Good Claude Code Prompt

A strong prompt has three parts:

1. What — the desired outcome
2. Constraints — what to preserve or avoid
3. Context — any non-obvious information Claude can't infer from the codebase

Example of a weak prompt:

"Fix the login"

Example of a strong prompt:

"The login form at src/app/(auth)/login/page.tsx throws a 401 on valid credentials when the email contains uppercase letters. Fix the comparison logic in the auth handler without changing the existing error message format."

Specificity Wins

The more specific your prompt, the more predictable the output. Vague prompts produce exploratory responses; specific prompts produce surgical changes.

If you don't know the file path, ask Claude first:

"Where is the user authentication logic? Read the relevant files and explain the flow."

Then follow up with a targeted edit request.

Iterative Prompting

Don't try to accomplish everything in one prompt. Claude Code excels at incremental work:

1. Ask Claude to explain the current state
2. Ask it to propose a plan
3. Execute one step at a time
4. Verify each step before continuing

Useful Prompt Patterns

Explain before edit: "Explain how the caching layer works, then propose a fix for the race condition."

Constrained edit: "Update the Button component to accept a loading prop. Don't change the existing prop interface."

Test-first: "Write a failing test for this edge case, then make it pass."