Prompt Engineer Skill

Help craft and optimize prompts for better AI model outputs.

Prompt Improvement Workflow

1. Analyze the current prompt
2. Identify weaknesses (vague, missing context, no format spec)
3. Apply techniques to improve it
4. Test with edge cases

Key Techniques

Role Prompting

Give the AI a specific role:

You are an expert {role} with {n} years of experience in {domain}.

Chain-of-Thought

Ask for step-by-step reasoning:

Think through this step-by-step before giving your final answer.

Few-Shot Examples

Provide input/output examples to guide behavior.

Output Format Specification

Define the exact format:

Respond in JSON with these fields: {field1}, {field2}, {field3}

Constraints & Guardrails

Set boundaries:

Keep your response under 200 words. Do not include {X}. Always include {Y}.

Prompt Evaluation Checklist

• [ ] Clear objective — what should the output be?
• [ ] Sufficient context — does the AI have what it needs?
• [ ] Output format — is the expected format defined?
• [ ] Examples — are there samples of good output?
• [ ] Constraints — are boundaries set?
• [ ] Edge cases — what could go wrong?

Guidelines

• Start simple, add complexity only if needed
• Test prompts with diverse inputs
• Prefer specific instructions over vague ones
• Use delimiters (quotes, XML tags) to separate sections
• Ask the user what model they're targeting