Reverse engineering an AI-powered game cheat

Armada researchers took apart a commercial AI aimbot and found a mature computer vision system built to evade the detection methods used against traditional cheats.

Cover of DestinyML Reverse Engineering Report

Download the full report to learn

  • How the cheat uses YOLOv5-family machine learning models to identify targets directly from screen pixels
  • Why screen capture, dual-PC configurations and hardware-like mouse inputs make traditional anti-cheat monitoring less effective
  • Which filesystem, process, network, USB and behavioral signals can help defenders detect this class of tooling
  • How AI-assisted reverse engineering helped Armada analyze the application, validate its behavior and develop actionable detection recommendations
Armada is a CREST accredited provider.

Get the report

We use your details to send what you asked for and occasional Armada research. Unsubscribe anytime. Privacy policy.

Trusted by security teams at

  • DISH
  • MapLarge
  • Workday
  • VMware
  • Salesloft

Start with the two-part blog series.

The blog posts cover the methodology and the headline findings. The report has the full technical notes.

Part 1: How AI agents changed the way we work

Why a security team reverse engineered a game cheat, and the agentic reverse engineering methodology behind it.

Part 2: What “AI aimbot” actually means

The findings on the cheat itself, and defensive takeaways for teams securing their own ecosystem.

CREST accredited

About Armada

Armada runs Continuous Threat Exposure Management. AI agents provide coverage and repetition, our team members provide judgment, and every finding ends in an action. Armada is a CREST accredited provider.

Get your copy