Getting Started
Quickstart
Run your first AI penetration test in under 5 minutes.
This guide walks you through launching your first pentest with CodeWall.
Step 1: Log in to the dashboard
Navigate to platform.codewall.ai and sign in with your credentials.
Step 2: Add a target
- Go to Pentests in the sidebar
- Click New Pentest
- Enter your target URL or IP address
- Select the target type (Web Application, API, MCP Server, or LLM Application)
Step 3: Configure scope
Define what CodeWall is allowed to test:
- In-scope domains — which hosts and paths the agent can interact with
- Out-of-scope exclusions — any endpoints or services to avoid
- Depth level — how aggressively the agent should test (see Safety & Guardrails)
If your target is behind a firewall or WAF, allowlist CodeWall's IPs before launching. See Configure Server Access for IP ranges and provider-specific instructions.
Step 4: Launch the test
Click Start Pentest. CodeWall's AI agent will begin:
- Reconnaissance — mapping the attack surface, discovering endpoints, and fingerprinting services
- Analysis — identifying potential vulnerabilities and planning attack chains
- Exploitation — safely exploiting vulnerabilities with proof-of-concept payloads
- Reporting — generating findings with full attack chains and remediation steps
Step 5: Review findings
Once the test completes, navigate to Findings to review discovered vulnerabilities. Each finding includes:
- Severity rating (Critical, High, Medium, Low, Informational)
- Detailed description of the vulnerability
- Step-by-step attack chain showing how it was exploited
- Proof-of-concept that your team can reproduce
- Remediation guidance
What's next?
- Core Concepts — understand targets, tests, findings, and more
- Test Types — explore the different kinds of tests CodeWall can run
- Safety & Guardrails — learn how CodeWall protects your systems during testing