# Choose your path
<a name="choose-your-path"></a>

Find the right starting point for what you want to do with AWS Security Agent. Use the table below to match your goal to a task, then go directly to that page. If you’re new to AWS Security Agent, read [What is AWS Security Agent?](what-is.md) first for a quick overview.


| I want to…​ | Who | Start here | 
| --- | --- | --- | 
| Understand what AWS Security Agent does before setting it up | Everyone |  [What is AWS Security Agent?](what-is.md)  | 
| Set up the service and create an Agent Space | Admin |  [Set up AWS Security Agent](setup-security-agent.md)  | 
| Test my live or deployed application for exploitable vulnerabilities | User |  [Quickstart: Run a penetration test](quickstart.md)  | 
| Scan my source code for vulnerabilities and policy violations | User |  [Quickstart: Run a code review](quickstart-code-review.md)  | 
| Model how my application could be attacked (STRIDE) | User |  [Quickstart: Run a threat model](quickstart-threat-model.md)  | 
| Get security feedback on a design before I write code | User |  [Create a design review](perform-design-review.md)  | 
| Scan code without leaving my IDE (Kiro or Claude Code) | Developer |  [Run code security scans from your IDE](code-review-ide-integration.md)  | 
| Scan only my changed lines before merging | Developer |  [Run a differential code scan with S3](run-diff-scan-s3.md)  | 
| Get automatic security comments on pull requests and merge requests | Admin |  [Enable code review](enable-code-review-scan.md)  | 
| Review findings and decide what to fix first | User |  [Penetration test](review-penetration-findings.md), [Code review](review-code-scan-findings.md), [Threat model](review-threat-model-findings.md), [Design review](review-design-findings.md)  | 

**Note**  
AWS Security Agent uses three roles, which you can identify by the interface you work in: **Admin** works in the AWS Management Console (setup and configuration), **User** works in the web application (running assessments and reviewing findings), and **Developer** works in GitHub or an IDE such as Kiro or Claude Code. A single person can hold more than one role. For full definitions, see [How AWS Security Agent works](how-it-works.md).