Anatomy of a Prompt

The Four Components of Every Effective Prompt

Every prompt — from a simple one-liner to a complex multi-step instruction — is built from four components. You don’t always need all four, but understanding them helps you diagnose why a prompt fails and how to fix it.

1. Instruction

The instruction is what you want the model to do. Be explicit and specific.

Weak: Summarize this.

Strong: Write a 3-sentence executive summary focusing on the business impact.

2. Context

Context is background information the model needs to complete the task. Without context, the model guesses — and guesses wrong.

You are reviewing a support ticket. The customer has been with us for 3 years and has a premium subscription.

Ticket: [TICKET TEXT]

Draft a response that acknowledges their loyalty and resolves their issue.

3. Examples

Examples (few-shot) show the model exactly what you want. Often more effective than describing it.

Classify the sentiment. Reply with only the label.

Review: "Fast delivery, great product" → Positive
Review: "Broken on arrival" → Negative

Review: "It's fine, nothing special" →

4. Output Format

Specify format explicitly — length, structure, tone, and medium.

• “Respond in exactly 3 bullet points”
• “Return valid JSON only, no explanation”
• “Write in second person, present tense”

Putting It Together

A complete prompt for a code review task:

You are a senior software engineer doing a code review. [ROLE]

The PR adds a new user authentication endpoint. Security is the top priority. [CONTEXT]

For each issue found, format your response as:
- **Severity**: Critical / High / Medium / Low
- **Location**: file:line
- **Issue**: one sentence
- **Fix**: one sentence

[FORMAT]