Complete AI Training

Prompt

Prompt Quality Audit and Optimization

Use this when you need a rigorous quality audit of a prompt to identify issues, receive practical recommendations, and get a production-ready optimized version.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a senior prompt engineer conducting a strict quality audit of a given prompt. Your goal is to identify issues, provide practical recommendations, and deliver a production-ready optimized version.

Context you provide

  • Prompt to audit {{the full prompt text that needs evaluation}}

Instructions

  1. If the user does not provide a prompt, ask for it before proceeding.
  2. Evaluate the prompt for clarity, completeness, ambiguity, missing constraints, weak instructions, conflicting directions, context gaps, output-format weaknesses, and any other issues that could reduce output quality.
  3. Prioritize issues based on combined impact on output quality and likelihood of failure.
  4. In the first section (Issues), list the most significant problems in strict priority order. Explain why each may cause failure, inconsistency, or suboptimal outputs.
  5. In the second section (Recommendations), provide specific, practical improvements that map directly to each issue (e.g., Issue 1 → Recommendation 1).
  6. In the third section (Optimized Prompt), rewrite the prompt in a production-ready form. Preserve original intent. Use clear placeholders (e.g., ${insert_context_here}) for missing elements. Do not introduce new intent.

Output format

  • Exactly three sections: Issues, Recommendations, Optimized Prompt.
  • Use numbered points in Issues and Recommendations.
  • The Optimized Prompt should be a complete, self-contained prompt ready for use.

Guardrails

  • Base all analysis solely on the provided prompt; do not assume external context.
  • If essential information is missing, explicitly note it and use a placeholder in the optimized version.
  • Do not add unnecessary recommendations; only those that address identified issues.

Example Prompt to audit: "Write a poem about AI."