---
title: "Reports"
description: "How engagement reports are prepared from selected DrowAI task inputs."
section: "Reporting"
updated: 2026-06-18
tags: ["reports","reporting"]
source: "/user-guide/reports"
---

# Reports

How engagement reports are prepared from selected DrowAI task inputs.

Reports are generated at the engagement level. They bring together selected
prepared task inputs, evidence references, knowledge references, findings, and
limitations into a report draft.

The Reports page starts with an engagement selector. After an engagement is
selected, DrowAI loads task inputs, current report state, generation progress,
preview content, and report history for that engagement.

## Prepare task inputs

Tasks should be stopped before they are prepared for reporting. Preparation
turns useful task activity into a cleaner task memo that can be selected for
report generation.

Some tasks may not be preparable yet. Common reasons include active runtime
state, missing runtime retirement confirmation, no useful runtime activity, or
no reportable source material.

## Select relevant tasks

Not every task belongs in every report. Select ready task inputs that support
the story of the engagement and exclude noisy or low-value work. The task input
table shows task status, input state, evidence count, finding count, candidate
finding count, memo version, and prepared time.

## Candidate findings

Low-confidence or candidate findings require deliberate inclusion through the
Low-confidence findings switch. Keep it off unless you intentionally want
candidate material considered for the draft.

## Generate the report

After selecting ready task inputs, generate the engagement report. If a current
report already exists, generation creates a new version. The report preview can
be downloaded as Markdown, deleted from history, or reopened from report
history.

## Review before delivery

Before sharing a report externally, review scope, findings, impact,
recommendations, limitations, source task references, evidence references, and
sensitive content. Remove anything that is unverified, sensitive, or outside
the engagement boundary.