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

# Reviewer agent

> How the Reviewer scores grant drafts against the evaluator rubric — checking word counts, citations, compliance flags, tone, and per-section quality.

<Note>
  **One job:** Block bad submissions before they go out.
</Note>

The Reviewer is the last agent in the grant pipeline. It receives the Writer's draft and scores it against the actual evaluator rubric for that specific call. Nothing is submitted until the Reviewer clears it.

## What Reviewer checks

| Check                 | Detail                                                     |
| --------------------- | ---------------------------------------------------------- |
| **Rubric alignment**  | Is each section answering what evaluators score?           |
| **Word count**        | Are sections within the call's limits?                     |
| **Missing citations** | Are claims backed by evidence from your Data Room?         |
| **Compliance flags**  | AI usage disclosure, EU AI Act gates, eligibility criteria |
| **Tone and format**   | Does the draft match the funder's expected style?          |

## Quality score

Every draft gets a quality score visible in the [Editor](/app/editor) dock. The score is broken down by section so you know exactly where to focus edits.

## Blocking submissions

If the draft doesn’t clear the Reviewer’s threshold, it is blocked from submission and a task is created in [Tasks](/app/tasks) with specific flags. You can:

1. Ask the Writer to fix the flagged sections
2. Edit manually in the Editor
3. Override the block (if your [autonomy level](/trust-and-control#autonomy-levels) allows)

<Warning>
  Submissions always require your explicit approval, regardless of autonomy level.
</Warning>