Cortex AES / Koi · Agentic Endpoint Security
One defender. One attacker campaign. Sixty seconds a call.
AES (Koi) Quest is a solo arcade range built on real classes of software and AI attack surface — malicious browser extensions, RCE-vulnerable packages, hijacked publishers, unregistered MCP servers. A simulated attacker campaign runs a meter against you every round. Read the evidence, make the call a strong security program would make, and take the high-score board.
- 12K unique items across 1,602 endpoints
- 35 malware detections on 277 endpoints
- 52% of the software estate governed
Findings are modelled on live Koi lab tenant data. The attacker campaign is a simulation scripted for this game — not live telemetry, not customer data.
How a run plays
3 acts · randomized rounds · covers the whole product over multiple runs.
Every run is reshuffled — a seed is stamped on each run.
Defender ranks
Earned per runYour final score sets your title on the high-score board.
Live lab access
Required for Act 2 · 3 live roundsAES (Koi) Quest runs alongside a live Koi lab tenant. Act 2 of every run is a guided discovery mission inside that tenant, and the optional Field missions add more. Every live round has an offline evidence fallback for half credit, so nothing dead-ends — but the real thing is the point. You need your own authenticated learning-center session before the launch link works.
-
1
Authenticate to the Palo Alto Networks learning center
Sign in with your own lab credentials / SSO. Do this first — the launch link fails without an authenticated session.
-
2
Open the live lab session
Launches straight into the Koi dashboard for the lab tenant.
Launch live Koi lablaunch.paloaltonetworks.com/palo-alto-networks/iocmetbdrsp45fsw2i67be
-
3
Acknowledge the cloned-environment notice
Land on Dashboard and keep the tab open: every scenario has a "Verify it live" pointer into Dashboard insights, Inventory, or Governance → Remediation.
What the game proves
Five outcomes a Koi buyer cares about, mapped to what you do on screen.
Lab tenant shape
Scenario metadata is modelled on these figures. Demo data — modelled on live Koi lab findings.
Risk categories in play
Handle a category well and you earn its badge.
Scenario packs
New run
Take the range
Pick a handle, choose your pack and run length, then go. Round 1 is a tutorial.
How scoring works
Round — of —
Arena
No run in progress
Start a run and the first scenario lands here with a simulated attacker campaign already moving.
Simulated attacker campaign
CAMPAIGN
Objective
Contain it before the meter fills. Simulation — scripted for this game.
Koi flagged this on your endpoints. Pick the response a strong security team would make — before the attack completes.
Your decision
Awaiting answerWhich Koi capability just saved you?
Debrief
Run summary
What you just did, translated into the outcomes a Koi buyer is funding.
Finish a run to see your debrief.
Global standings
High scores
Untimed · live Koi tenant
Field missions
Untimed. Open the live Koi tenant and find the answer — this is the product doing for real what the game simulated.
Secondary · optional
Facilitator tools
Running this in a room? Everyone plays their own solo run on their own device. These are the few controls a facilitator needs.