Articles
Go back to application
Articles
Go back to application
Platform Fundamentals
Getting Started
Account Management
User Management
AI Pentest
Getting started
Initiate AI Pentest
Workspaces
Templates
Configuration
Operations
Asset Management
Asset Basics
Assets Operations
Asset Customization
Vulnerability Management
Vulnerability Basics
Vulnerability Operations
Vulnerability Organization
Integrations & Connectors
Credential Management
Connector Setup
Synchronization
Scanning and Assesssment
Scanning Operations
Security Analysis
Workflow Configuration
Engagemet Management
Enagagement Basic
Enagagement Operations
Workbook
Pentesting as a Service
PTaaS Basics
Vulnerability Submission
Task Management
Task Basics
Task Automation
Reporting and Documentation
Reports
Documentation
Attack Surface Management
Strobes Attack Surface Management (ASM)

Initiate AI Pentest

This article walks you through the high-level flow of running your first AI-driven penetration test on Strobes

Every AI pentest on Strobes follows the same five-action loop:

  • Open the AI Workspaces page
  • Create a new Workspace
  • Pick a template (or a Blank Workspace)
  • Fill in target, credentials, model, and (optionally) a schedule
  • Launch the assessment

Now, let's take a look at each of these actions.

1. Open the AI Workspaces page

From the left-hand navigation, open AI → Workspaces. You'll land on the workspaces list, which shows all active and archived assessments for your organization.

2. Create a new Workspace

Click the New Workspace button at the top right (or the + Create Workspace card if the list is empty). This opens the Create Workspace wizard.

3. Pick a template or a Blank Workspace

The wizard's first step is Template. Pick a pre-configured template (e.g., Web App Pentest) for a standard assessment with a defined methodology, or select Blank Workspace for a free-form, chat-driven run.

4. Fill in target, credentials, model, and schedule

Move through the Scope and Configure steps to define what the agent should test:

  • Target — a URL, IP range, repository, or cloud account, depending on the template.
  • Credentials (optional) — test accounts, API keys, SSH keys, or cloud roles the agent may use.
  • AI Model — the LLM driving the agent (Standard, Advanced, or Lite).
  • Schedule — One-time, Daily, Weekly, Monthly, or a custom cadence.

5. Launch the assessment

On the Review step, double-check your choices and click Launch Assessment (or Create & Launch). The AI agent takes over from there, and you can watch progress on the workspace dashboard.

Was this article helpful?

Powered by Product Fruits