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

# Overview

> agentgg is a command-line security scanner. It reads your code and confirms a finding before it reports it.

A pattern matcher flags a line and moves on. An agentgg agent reads the code around that line, follows the imports, and traces who calls it. It reports the finding only when the code confirms it.

<Card title="agentgg on GitHub" icon="github" href="https://github.com/agentgg-dev/agentgg">
  The CLI is open source under the Apache 2.0 licence. Read the code, open an issue, or send a pull request.
</Card>

## The report

A scan writes a report to an output directory of your choice. The report has
two parts.

* **A summary.** One page with the counts, the findings grouped by agent and by
  severity, and a link to each finding.
* **One file per finding.** Each names the source file and the exact lines, then
  explains the problem, how to reproduce it, and what it costs you.

Every finding carries a severity, a CVSS score, and a verdict that says whether
a second pass confirmed it. See [Resume and reports](/cli/resume-and-reports)
for the full shape of both parts.

## What you need

* Node.js 20 or later.
* An account with one model provider. See [Providers](/cli/providers) for the list of providers.

## The hosted platform

The platform runs the same agents on our machines instead of yours. It needs no
install, and it provides the model, so you bring no API key.

<Card title="The agentgg platform" icon="cloud" href="/platform/overview">
  Scan from a dashboard, scan every pull request, and share findings with your team.
</Card>