Rule reference · Clarity

Long sentences

  • long_sentence
  • info in this example
  • Clarity
  • stage · Instruction quality

Fires when: A prose sentence over 40 words.

One long sentence often carries four or five requirements joined by "and" and "while". Each is easy to overlook, and it is hard to tell which one the model skipped.

Put one requirement per sentence or bullet. It also makes the prompt easier to review and diff.

Example that triggers it

Review the pull request below and point out any correctness problems, security issues such as injection or missing authorization checks, and places where the code is hard to maintain because of duplication or unclear naming, while ignoring purely stylistic preferences and suggesting a concrete fix for each problem you report.

info 1 sentence over 40 words.

Why it matters. Long sentences bundle several instructions together; models are more likely to miss one of them.

Fix. Split each requirement into its own sentence or bullet.

Quality 84/100 · 55 tokens · 3 findings

Fixed version

Review the pull request below.

- Report correctness problems.
- Report security issues, such as injection or missing authorization checks.
- Report maintainability problems, such as duplication or unclear naming.
- Ignore purely stylistic preferences.
- Suggest a concrete fix for each problem.

Return a Markdown list grouped by category.

<diff>
{{diff}}
</diff>

Quality 100/100 (+16) · 70 tokens · 0 findings

Resolved by the fix

  • missing_output_format: No output format is specified.
  • unclear_reference: The prompt refers to something ("this", "the above", "the file") that is not included.
  • long_sentence: 1 sentence over 40 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 Clarity dimension. See the scoring model for the full formula.

All 20 rules →