> ## Documentation Index
> Fetch the complete documentation index at: https://docs.fibonax.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Quality score

> SprintBrain scores each prompt out of 10, shows what's missing, and fixes most gaps in one click.

The prompt editor scores the prompt out of 10 while you write:

* **8 or more**: Excellent.
* **5 to 7.9**: Good.
* **Below 5**: Needs improvement.

Open the score to see each check, why it passed or not, and an example of what would pass.

## What it checks

| Check | Passes when |
| - | - |
| Role defined | The Role block is on and filled in. |
| Clear objective | The Objective block states the task. It counts the most. |
| Context provided | The Context block has background. |
| Constraints defined | The Constraints block sets limits. |
| Examples included | The Examples block shows sample input and output. |
| Audience specified | The prompt names who the answer is for. |
| Self-check requested | The prompt asks the AI to check its work or flag doubts. |
| Output format set | An output type is chosen. |
| Strategy selected | A strategy is chosen. |
| Sufficient depth | The whole prompt is detailed enough to be clear. |
| Model selected | A preferred AI model is set. |

A check can pass in part, for example a block that's on but thin.

## Fix it in one click

A check that doesn't pass shows a button with the fix, such as **Enable** a block, **Name the audience**, **Add self-check** or **Apply One-shot**. Click it and SprintBrain makes the change for you.

## Compare with your library

When you have enough prompts, the score also says how this one compares, for example "Stronger than 68% of your 12 prompts".

<Note>
  The score measures structure, not results. A prompt you've tested on real work may beat a higher-scoring one.
</Note>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.