> For the complete documentation index, see [llms.txt](https://docs.ox.security/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.ox.security/scan-and-analyze-with-ox/scanning/agentic-pentester-new.md).
# Agentic Pentester
> **Note:**
>
> * This capability is currently in Early Access (EA) and is not generally available. To request access, please contact OX technical support.
## About Agentic Pentester
Agentic Pentester enables you to perform automated penetration testing against web applications using AI-driven agents.
Unlike traditional security scanners that rely on predefined checks, Agentic Pentester performs real attacks against live targets. The agent explores the application, discovers reachable endpoints, identifies attack paths, executes attack techniques, and validates findings with evidence.
Agentic Pentester combines runtime discovery with source code intelligence. By correlating discovered endpoints with connected repositories, OX can better understand application behavior, improve attack accuracy, and provide richer investigation context.
Use Agentic Pentester to:
* Configure applications for assessment.
* Verify that OX can access and authenticate to targets.
* Discover exposed URLs, APIs, and attack surfaces.
* Identify vulnerabilities across the application.
* Correlate runtime findings with source code repositories.
* Visualize relationships between targets, endpoints, and findings.
* Review attack evidence and understand how vulnerabilities were discovered.
| Item | Description |
| ---------------------- | -------------------------------------------------------------------- |
| Target Name | Name assigned to the target. |
| Asset Summary | Displays the number of Web and API assets discovered for the target. |
| Deployment Environment | Development, Staging, or Production environment. |
| Connectivity Status | Current connectivity verification status. |
| Findings | Number of findings grouped by severity. |
| Last Scan | Indicates when the target was last assessed. |
### How Agentic Pentester Works
Agentic Pentester analyzes applications from two directions:
* Runtime discovery of the running application.
* Source code intelligence from connected repositories.
The agent discovers exposed endpoints, explores application behavior, identifies attack paths, executes security tests, and validates findings using evidence collected during the assessment.
By combining runtime discovery with repository intelligence, OX creates a unified view of the application and its attack surface.
The typical Agentic Pentester workflow is:
1. [Create a target and connect repositories](/scan-and-analyze-with-ox/scanning/agentic-pentester-new/adding-targets.md#step-1-create-a-target-and-connect-repositories)
2. [Verify connectivity](/scan-and-analyze-with-ox/scanning/agentic-pentester-new/adding-targets.md#step-2-verify-connectivity)
3. [Ensure target access](#ensure-target-access)
4. [Run an assessment](#run-assessments)
5. [Review discovered endpoints and findings](#review-agentic-pentester-findings)
6. [Review target details](/scan-and-analyze-with-ox/scanning/agentic-pentester-new/reviewing-findings.md)
7. [Export a penetration testing report](/scan-and-analyze-with-ox/scanning/agentic-pentester-new/exporting-a-pen-test-report.md)
## Ensure target access
Before assessing protected applications, ensure that OX can access the target.
Some applications are protected by Web Application Firewalls (WAFs), network firewalls, IP allowlists, or other security controls that can block assessment traffic.
To allow Agentic Pentester to access the application, allow traffic from the following OX static IP addresses:
* 18.202.47.201
* 63.33.147.64
If the application uses CAPTCHA or a custom access mechanism, configure a custom header when creating the target.
## Run assessments
Agentic Pentester scans run independently from the main OX scan.
Currently, Agentic Pentester scans can be triggered manually only.
The following scan options are available:
| Option | Description |
| ---------------- | -------------------------------- |
| Scan All Targets | Scans all configured targets. |
| Scan n Targets | Scans only the selected targets. |
When a scan starts, OX displays scan progress and the number of targets currently being assessed.
Agentic Pentester findings are validated through successful attacks against live targets, providing a higher level of confidence than findings based solely on static analysis or configuration inspection.
**To start a Pentester scan:**
1. From the upper right corner, select **Scan All Targets** or **Scan \ Targets**. The progress bar appears, with the details that include how many targets are being scanned.
2. To view the Agentic Pentester issues on the **Active Issues** page, go to **Agentic Pentester > Issues**.
## Review Agentic Pentester findings
Agentic Pentester findings are displayed on the Active Issues page and include an additional Attack Overview tab that provides a step-by-step record of how the vulnerability was identified and validated.
The Attack Overview tab includes:
| Section | Description |
| --------------- | ------------------------------------------------------------------------------------- |
| Attack Recap | High-level summary of the attack and validation process. |
| Attack Timeline | Chronological view of the actions performed by the agent. |
| Request Details | Request parameters, headers, payloads, and target information used during the attack. |
| Evidence | Artifacts collected from the target and used to validate the finding. |
Use the Attack Overview tab to understand how the vulnerability was discovered, what evidence was collected, and why the finding was generated.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://docs.ox.security/scan-and-analyze-with-ox/scanning/agentic-pentester-new.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.