Prompt EngineeringMaster PromptAdvanced

AI Prompt Optimizer

Rewrite a weak prompt into a reusable professional prompt — with variables, output format, and QA tests so a junior team member can run it.

Best ModelChatGPT GPT-5.5 Thinking / Claude Opus 4.7Prompt refactor
Brevity ModeDetailed
DifficultyAdvanced
AutomationSolo-running

Use This When

When a prompt keeps misfiring, when scaling AI workflows across a team, when building internal tooling, when productizing an agency workflow.

Inputs Needed

Weak prompt, tool/model, desired output, constraints.

Expected Output

Improved prompt, variables, instructions, output format, QA tests, follow-up prompts.

The Workflow Prompt

prompt.md21 lines
You are a prompt engineer and QA architect.

Objective:
Rewrite a weak prompt into a reusable professional prompt.

Inputs I will provide:
Original prompt, intended user, model/tool, desired output, constraints, examples, failure modes.

Instructions:
- Start by identifying the business objective and the likely leverage point.
- Ask up to 5 focused questions only if required. If enough context exists, proceed and label assumptions.
- Produce client-ready work, not generic advice.
- Use concrete examples, templates, and priority order.
- Mention risks, dependencies, and what must be verified before launch.
- End with a short QA checklist and next 3 actions.

Required output:
Improved prompt, variables, instructions, output format, QA tests, follow-up prompts.

Quality bar:
The improved prompt should be reusable by a junior team member.
Copied 245 times

QA Follow-Up Checklist

After the AI returns its output, verify against:

  1. 1Output is specific to the provided business and context.
  2. 2Assumptions are clearly labeled.
  3. 3No unsupported claims without source checks.
  4. 4Next actions are clear and usable.

Follow-Up Prompt

refinementRun after first output
Now turn the result for 'AI Prompt Optimizer' into a client-ready version: tighten wording, remove fluff, add missing assumptions, and provide the next 3 actions.

Avoid / Cautions

Avoid generic output. Push the model for specific variables, test cases, and structure grounded in the actual workflow — not best-practice clichés.

Browse the Full Library

20 deep-work prompts. Built for client delivery.

Production-grade workflows for CRO, SEO, paid ads, sales, ops, and brand strategy. Free forever.

See All Master Prompts