ForkFree

Prompts / Writing

Expand a vague draft into a clear brief

It turns a short, half-formed request into a complete instruction with a goal, a format and boundaries. This works because the AI no longer has to guess those three things, so you spend fewer rounds rewording.

Partly workedIt kept the original sentences, gave a clear goal, format and boundaries, included the earlier context and exclusions, and used [assumed] tags with no commentary. But it isn't "short": it has a long nested list. Several guesses aren't marked [assumed], including the column set, the console summary, the code-block and setup-instructions requirement, and the logging. One assumed tag ("chosen by you") is muddled.

The prompt

Act as a careful prompt editor who turns rough requests into instructions an AI can carry out first time.

Task: Rewrite the draft request below so it states a clear goal, an output format and sensible boundaries, without changing what I asked for.

Draft request: {draft_request}
Earlier context I have already settled: {confirmed_context}
Things to rule out: {ruled_out}

Steps:
1. Keep my original sentence as the opening line, unchanged.
2. Add a specific goal: name the exact items, quantity and purpose you can reasonably infer.
3. Add the output format: file type, structure, length or layout.
4. Add boundaries: limits, safety or politeness rules, and anything from the earlier context or the ruled-out list.
5. Mark any detail you had to guess with [assumed] so I can correct it.

Output: Only the rewritten request as one short paragraph or a short bulleted list, ready to paste. Add no commentary before or after it.

Change the parts in curly braces. Use it before sending any one-line request to a chat model, especially for coding or research tasks.

What happened when we ran it

We sent

Act as a careful prompt editor who turns rough requests into instructions an AI can carry out first time.

Task: Rewrite the draft request below so it states a clear goal, an output format and sensible boundaries, without changing what I asked for.

Draft request: Help me write a script that collects the prices of running shoes from a few online shops so I can compare them. I run a small sports store and want to keep our prices competitive.
Earlier context I have already settled: I use Python on Windows, and I only care about men's road running shoes sold in the UK in pounds.
Things to rule out: No paid tools or subscriptions, and no logging in to any accounts.

Steps:
1. Keep my original sentence as the opening line, unchanged.
2. Add a specific goal: name the exact items, quantity and purpose you can reasonably infer.
3. Add the output format: file type, structure, length or layout.
4. Add boundaries: limits, safety or politeness rules, and anything from the earlier context or the ruled-out list.
5. Mark any detail you had to guess with [assumed] so I can correct it.

Output: Only the rewritten request as one short paragraph or a short bulleted list, ready to paste. Add no commentary before or after it.

The model replied

Help me write a script that collects the prices of running shoes from a few online shops so I can compare them. I run a small sports store and want to keep our prices competitive. - **Goal:** Write a Python script (Windows) that collects the product name, brand, current price in GBP, and product URL for men's road running shoes from 3 UK online shops [assumed: 3 shops, chosen by you from well-known UK running or sports retailers, with the shop list kept in an easy-to-edit config section at the top]. Its purpose is to let me compare competitor prices against my own and keep my store's prices competitive. - **Output format:** Save results to a single CSV file [assumed: `shoe_prices.csv`, UTF-8, overwritten each run] with one row per shoe and these columns: shop, brand, model, price_gbp, url, date_collected [assumed: ISO date format]. Add a short summary printed to the console showing the number of shoes collected per shop. Provide the full script in one code block, followed by brief setup and run instructions (no more than 5 lines). - **Boundaries:** - Scope: men's road running shoes only, UK shops only, prices in pounds only. Skip trail, track, kids' and women's products, and any non-GBP prices. - Cost and access: use only free tools and libraries [assumed: `requests` and `BeautifulSoup`, installable with pip], with no paid services or subscriptions and no logging in to any account. - Politeness and safety: check each site's robots.txt and terms and skip any shop that disallows automated collection, add a delay of at least 2 seconds between requests [assumed], send an honest User-Agent, collect only a limited number of pages per shop [assumed: up to 5 listing pages], and do not attempt to bypass CAPTCHAs, blocks, or rate limits. - Reliability: handle missing prices or failed requests without crashing, and log any skipped items or shops to the console. - Keep the code simple and commented so a non-expert can adjust the shop list and selectors.

Run on Claude Sonnet 5.5, 08 Nov 2026. Automatic check: 3 out of 5.

Mentioned for

DeepSeek, ChatGPT, Claude

Where it came from

Found on GitHub, shared by huaguoxingkong: huaguoxingkong/dsh-prompt-rewriter.

This is our own version of the idea, written and tested by us. The original may differ.

More writing prompts

Writing Struggled in our test
Rewrite Machine-Sounding Text Without Inventing Facts

You get a rewrite that sounds like a person wrote it, with a clear list of what was changed. The locked fact list stops the model inventing statistics or stories to seem more human.

Found on GitHub, shared by yucezerey
Writing Worked in our test
Add One Missing Requirement to Drafts

You get one targeted addition to your prompt instead of a full rewrite, with a reason for it. It works because the AI must first list what is already good, so your original wording and constraints are preserved.

Found on GitHub, shared by dbcaskml183-dotcom
Writing Worked in our test
Stretchtext Replies With Expandable Detail

You get a short answer that grows in controlled layers, with each new addition highlighted so nothing is repeated or lost. It works because each level is forced to keep the earlier words, so the reader stays oriented as detail is added.

Found on GitHub, shared by pro-vi