The scope is the list of targets an AI Network Pentesting assessment may touch. This page covers what you enter, the limits that apply, and which ports Escape scans on each target.

## Accepted Targets

Enter one target per line, or import a CSV.

| Target | Example | Notes |
| --- | --- | --- |
| Domain | `edge.example.com` | Escape resolves its DNS records and tests the hosts behind it. |
| IPv4 address | `192.0.2.10` |  |
| IPv4 CIDR | `192.0.2.0/24` | Stays one range: it isn't split into individual targets. |
| IPv6 address | `2001:db8::1` | Single addresses only. |

Escape normalizes CIDR ranges to their network address, so `192.0.2.15/24` is saved as `192.0.2.0/24`.

## Rules and Limits

- **At least one domain or single address.** A scope made only of CIDR ranges is rejected. The first domain or single address is the profile's main host.
- **Ranges up to `/24`** (256 addresses) by default. Wider ranges, up to `/16`, are available on request: contact your Escape representative.
- **No IPv6 ranges.** Enter IPv6 addresses one by one.

Invalid lines are listed under the scope field for you to fix or remove.

## Out-of-Scope Hosts

Agents only attack the targets in your scope. When agents come across a host outside it, Escape skips that host and records it in the assessment logs as a skipped out-of-scope host.

The **Context** fields in **Fine-Tune (Optional)** narrow what agents focus on. They never add targets to the scope.

## Ports Scanned

Escape scans TCP ports with a connect scan, then runs protocol detection on every open port.

| Target | Ports scanned |
| --- | --- |
| Domain or single address | 3,846 TCP ports, plus any port seen on the host before. See the [full default list](<https://docs.escape.tech/documentation/asm/network-scanning/#port-scanning>). |
| CIDR range | 9 ports: 22, 80, 443, 3306, 5432, 6379, 8080, 8443, 27017 |

Range scans probe every port on every address, so the default set is small. To cover more ports on a specific machine, add its address as a single target as well.

Every open port gets its own reconnaissance pass before agents attack it. Reconnaissance doesn't send attack payloads.

## Validation

**Validate Configuration**, in the **Review &amp; launch** step, checks every target from the selected location. It doesn't start the pentest, and **Create &amp; run assessment** stays disabled until it passes.

## Location

You choose where traffic comes from in **Fine-Tune (Optional) \> Location**: the Europe, United States, or Canada region, or one of your [private locations](<https://docs.escape.tech/documentation/platform/private-location/index.md>).

- **Public region**: tests what the internet sees. Use it for public IPs and domains.
- **Private location**: tests from inside your network. Use it for private ranges such as `10.10.5.0/24` or `192.168.50.0/24`, and for internal domains that only resolve on your network.

The source IP addresses for each public region are listed on [Firewall Configuration](<https://docs.escape.tech/documentation/platform/firewall/#public-locations>).

## Related

- [Attack Vectors](<https://docs.escape.tech/documentation/network-pentesting/attack-vectors/index.md>): what agents may attempt on the targets in scope
- [Guardrails](<https://docs.escape.tech/documentation/network-pentesting/guardrails/index.md>): rate limit, duration, and stopping an assessment
- [Network Scanning in ASM](<https://docs.escape.tech/documentation/asm/network-scanning/index.md>): the same port lists in your ASM inventory