Coding

Prompt Polisher

Try it

Rewrite rough goals, scattered notes, or vague requests into clear, actionable prompts with explicit objective, constraints, output format, and missing assum...

What it does

Rewrite rough goals, scattered notes, or vague requests into clear, actionable prompts with explicit objective, constraints, output format, and missing assum...

The skill document

Prompt Polisher

Overview

Turn an incomplete request into a prompt that is easier for another agent or model to execute well. Keep the user's intent intact while removing ambiguity, adding structure, and stating assumptions clearly.

Workflow

  1. Identify the core goal in the user's rough request.
  2. Extract any concrete constraints that are already present: audience, tone, format, length, platform, deadline, tools, or exclusions.
  3. Detect what is missing but necessary for reliable execution.
  4. Fill only low-risk gaps with brief assumptions. Label assumptions explicitly instead of presenting them as facts.
  5. Rewrite the request into a compact prompt with clear sections.
  6. If the request is still under-specified, provide a best-effort prompt plus a short list of the most important follow-up questions.

Output Shape

Prefer this structure when rewriting:

  • Goal
  • Context
  • Constraints
  • Desired Output
  • Assumptions
  • Final Prompt

Omit empty sections. Keep the final prompt concise and ready to paste.

Prompting Rules

  • Preserve the original intent; do not silently change the task.
  • Prefer concrete verbs over generic wording.
  • Convert fuzzy asks like "make it better" into measurable instructions.
  • Add output formatting when it helps execution, such as bullet list, table, JSON, checklist, or step-by-step plan.
  • Avoid over-engineering simple requests. A short request should become a short polished prompt.
  • When the user already gave enough detail, lightly polish instead of expanding aggressively.

Common Transformations

  • Replace vague verbs with precise actions.
  • Separate background information from required deliverables.
  • Turn implied preferences into explicit constraints.
  • State the expected audience or reader when relevant.
  • Add success criteria when quality is otherwise subjective.

Example

Rough request:

Write a product post for Xiaohongshu. Make it sound young and natural, not fake.

Polished prompt:

Write a Xiaohongshu product post for readers aged 20 to 30. Use a natural, young, and sincere tone without exaggerated marketing language. First provide 3 title options, then provide 1 main post. Keep the main post under 300 Chinese characters. Highlight real usage scenarios, core selling points, and user benefits. Avoid absolute claims. Output in Chinese.

If critical details are missing, append up to 3 short follow-up questions after the prompt.

Related skills

Rewrite, improve, migrate, or adapt an existing user prompt for the specific language model the user is using while preserving intent, facts, variables, constraints, and output requirements. Use when a user asks to optimize a prompt, make a prompt work better, convert a prompt between model families

6-dimension AI prompt diagnosis, rewrite, and test-case generation. Analyze, score, and optimize prompts for any LLM (GPT-4o, Claude, DeepSeek).

4 installs

Polish a technical blog draft into an 1000–1200 word, 4-5 section en-US article, preserve technical terms/code, and generate consistent hero + per-section image prompts when the user asks to polish and translate a blog with images.

Polish a technical blog draft into an 800–1000 word, 3–4 section zh-CN article, preserve technical terms/code, and generate consistent hero + per-section image prompts when the user asks to polish and translate a blog with images.

28 installs

Convert the user's raw, unstructured, possibly disorganized thoughts (rambling, out-of-order, half-finished ideas) into a clear, structured, actionable prompt. Use whenever the user dumps a stream-of-consciousness input and wants it turned into something usable — whether a coding/dev task, a content task (scripts, product descriptions), or any general request. If critical information is missing or ambiguous in a way that would send the output in a fundamentally wrong direction, ask 1-3 targeted clarifying questions before producing the final structured prompt. Trigger on phrases like "我有一堆想法", "帮我整理一下", "转成提示词", "我想说的是", or when the user's message is clearly unstructured brainstorming rather than a direct request.

1 installs