---
title: "Overview"
description: "The main product terms used across the DrowAI guide."
section: "Foundations"
updated: 2026-06-18
tags: ["concepts","terminology"]
source: "/user-guide/overview"
---

# Overview

The main product terms used across the DrowAI guide.

DrowAI is easier to understand when its main objects are separated clearly. The
product is not a single automatic scan button. It is a workspace for directing
authorized testing, observing what happens, preserving useful material, and
preparing report drafts.

## Engagement

An engagement is the top-level container for related work. It can represent a
client assessment, an internal review, a red-team exercise, or a lab/demo
scenario. Tasks, knowledge, and reports stay attached to the engagement that
owns them.

## Task

A task is a scoped work session. When a task is created, it is attached to an
engagement, given a workspace, and prepared for runtime work. The operator can
start, pause, resume, stop, chat about, inspect, and eventually prepare task
activity for reporting.

## Scope

Scope defines the task boundary. It can be typed into the task form or loaded
from a text/Markdown scope file. Scope may include targets, objectives,
constraints, exclusions, and other context. Public examples should use mock
targets or approved lab environments only.

## Evidence

Evidence is durable material saved from task activity. It may include tool
output, artifacts, excerpts, metadata, or other source material that supports
later review. Evidence is useful because it lets the operator trace a finding
or report claim back to observed source material.

## Finding

A finding is a security-relevant observation associated with an asset, service,
or activity. DrowAI can organize findings and candidate findings, but a human
operator should review them before treating them as final report content.

## Knowledge

Knowledge is DrowAI's durable view of observed work: findings, assets,
services, evidence, and relationships. The Knowledge workspace is for review
and navigation, not for hiding the need for human judgment.

## Report

A report is an engagement-level draft generated from selected prepared task
inputs. A report can have versions, history, source task references, evidence
references, and downloadable Markdown. It should be reviewed and edited before
external delivery.