Autoattack

Your goal directs the attack.

Give the agent a concrete goal, such as administrator access to a backup server or a read of a protected document. AutoAttack chooses attacks that can advance it toward that result.

You decide what matters

Administrator access to a backup server and a read of one file require different privileges. Name the result you want the agent to pursue. Leaving the objective blank directs it toward Domain Admin on every reachable domain.

The objective guide works through that choice using a backup-server example.

Each foothold opens another attack

A password exposed in a file may open another server. Access to that server may reveal a more privileged account. AutoAttack chains these gains at machine speed, using each result to pursue the next attack toward your goal.

Our published experiments show how that approach played out in a lab. The Platform page explains how to give a campaign its starting access.

Make the result checkable

A claim of administrator access needs evidence of that privilege. The report connects access gains to the captured output that supports them, and subscribers can inspect it before acting on the recommended fix.

The report states whether the campaign reached the requested access and shows the gains that advanced it toward that objective.

Keep the context for the next decision

The system owner needs to know what allowed access and which dependencies a change might affect. Keeping the route and remediation advice together makes that conversation specific.

Every follow-up campaign starts fresh. Set its objective and scope to test the systems or permissions you changed.

Give the agent a goal.

Set the objective and authorized scope, then follow the access the campaign gains.

Contact
Autoattack