Rule reference · Specificity
Prompt too short
Fires when: Fewer than 5 content words.
A four-word prompt is not wrong, but every missing decision is made by the model: topic, length, tone, audience, and format. The result is generic and changes on every run.
Add only what changes the output: the subject, who it is for, a length, and the format. That usually takes one sentence, not a page.
Example that triggers it
Write a poem.
warning The prompt is only 3 words.
Why it matters. Very short prompts leave scope, audience, and format to the model.
Fix. Add the context, constraints, and the output you expect.
Quality 87/100 · 4 tokens · 2 findings
Fixed version
Write a 12-line poem about autumn in a small fishing town, for adult readers. Return plain text.
Quality 97/100 (+10) · 22 tokens · 1 findings
Resolved by the fix
too_short: The prompt is only 3 words.
Scores, token counts (GPT-4.1 tokenizer), and findings on this page are computed by the engine when the site is built.
How it affects the score
Each finding subtracts a fixed penalty from 100: critical 30, warning 10, info 3. This rule counts against the Specificity dimension. See the scoring model for the full formula.
Other specificity rules
- Unquantified size: "short", "a few", "detailed" with no number.
- Missing programming language: Code requested without a language or framework.