Rogue Agent Console
Security control simulator

Rogue Agent Console

Set your Essential Eight maturity and a second set of agent-specific controls on the left. Pick a scenario, then see how far a misbehaving research agent gets, how quickly it would be detected, contained and recovered from, and what that means across a decentralised organisation. Built for teaching, not a measurement of any real system.

Environment preset
Scenario

Essential Eight

ACSC's baseline mitigation strategies. Maturity Level 0 (not implemented) through Level 3.

Agent-specific controls

Nine safeguards the Essential Eight was never designed to cover. They apply to autonomous agents acting on your behalf, including when the agent has legitimate access.

Key takeaway

Attack chain, current configuration

Same scenario with Essential Eight alone (agent-specific controls set to zero):

Same scenario with agent-specific controls alone (Essential Eight set to zero):

Detect, contain, recover

If the agent does get going, how long before someone notices, how long to switch it off and cut its access, and how long to get back to a clean state. Times are illustrative and depend on the logging, containment and backup settings.

Maturity at a glance

The Essential Eight is rated at the lowest level reached across its eight controls, which is how the ACSC model works. The agent-control scale is this tool's own and is not an ACSC measure.

Three environments side by side

The same scenario run against no meaningful controls, the Essential Eight at Maturity Level 3, and the Essential Eight plus agent-specific controls at their highest level.

A decentralised estate

Large universities run many units with their own IT, budgets and risk appetite. Choose a profile for each unit and see the same scenario play out across the estate.

Network path

Where the agent session can actually reach, given your current settings. Numbers match the stages in the attack chain.

Agent session 3 6

Session log

What the agent actually did at each stage, in plain terms.