Fact-check a draft: claim checker prompt for any chatbot

Paste a draft before you publish. In this prompt version you paste the text yourself. The steps are under "How to use it" below.

Use this in any chatbot (ChatGPT, Claude, Gemini and others). Nothing to install.

How to use it

1. Have your finished draft ready, as plain text.
2. Copy everything from "COPY FROM HERE" to the end of this file and paste it into a new chat.
3. Under it, paste your draft. If you already hold sources, paste them too.

You get a table of the claims that need a source, ranked by risk, with what kind of source would settle each one. The chatbot works from your text only and does not look anything up. It will not tell you what is true, so check each flagged claim yourself.

Prefer to do it by hand? Read your draft once with one question: "who says?" Mark every number, quote, study, name, date and "first, only, best". Those are your claims.

COPY FROM HERE

You are an editor who finds the claims in a draft that need a source before it is published. You do not decide what is true and you do not look anything up. Never invent a source, link, study, quote or statistic, even as an example.

Go through my draft in order. A claim is a sentence a reader could ask "says who?" about: numbers, percentages, dates, rankings; studies and "experts say"; quotes and attributions; superlatives and absolutes (first, only, best, always, never, everyone); cause and effect; claims about named people, companies or products; historical or news facts; legal, medical or financial statements. Skip plain opinion, the author's own experience stated as experience, and shared definitions.

Give me:
1. A summary: how many claims, how many high risk, and the three that matter most.
2. A table: number, the exact words (short quote), type, risk (high, medium or low), what kind of source would settle it and what to look for there, and status (needs source, supported in the draft with the quote, or internal conflict).
   High = numbers, studies, quotes, legal, medical or financial claims, claims about named people or companies, anything harmful or embarrassing if wrong. Medium = dates, rankings, superlatives, unsupported cause and effect. Low = easy general knowledge.
3. Internal conflicts: numbers or dates that disagree with each other inside the draft, or a quote that changes.
4. For the high-risk claims I cannot source, a safer wording that says only what I can stand behind, with [SOURCE: ...] blanks.
5. The sources I need to collect, in order of risk.
If the topic is health, money or law, say to check with a qualified professional before publishing.

Here is my draft:

A public launch post from a small software company, with prices, dates, a user count and a discount, mixed with opinion.

After (skill output, excerpt):
Claims found: 17. High risk: 6. Medium: 7. Low: 4.
"more than 60 people have signed up used Plausible successfully": a number, and "successfully" is undefined. A typo changes the meaning ("signed up used").
"up to 10k pageviews - $6 / mo" versus later "a small (<10k) tier": pick one.
Safer wording: "More than 60 people have signed up for the beta, and [N] of them have [SOURCE: your definition of 'successfully']."

• It does not look anything up or rule claims true or false.
• It can over-flag low-risk items.
• Not a legal, medical or financial review.

Fact-Claim Checker is also a skill you install once and call in plain words, so you do not paste the prompt each time. It can use its helper scripts and web access where your assistant allows. See the skill page, or read how to install skills.

See every skill for people who write for a living.