T-70: Facial Verification Malware POC

Project Overview

T-70 is a Rust project inspired by the Cyberdyne Series 70 Automated Fighting Infantry Unit ("The Terminator"). The project performs runtime facial verification using OpenCV through the opencv-rust bindings and integrates a CascadeClassifier with MobileNetV2 models for face detection and similarity comparison. The system is designed to run as a standalone executable, making it ideal for deployment in environments where real-time facial verification is essential.

Key Features:

• VM Detection: The project includes virtual machine detection logic to determine whether the executable is running in a virtualized environment. If a VM is detected, the executable self-deletes.
• Face Detection and Comparison: The system captures images using a connected camera, detects faces, and compares embeddings from the live feed with a reference image to compute similarity scores.
• Automated Action: Based on the similarity score, the system can run shellcode if a match is found.

Project Structure

• main.rs: The main entry point of the project.

• Modules:

  • camera: Handles camera initialization and input.
  • classifier: Manages the initialization of the face detection classifier.
  • face_detection: Responsible for detecting faces and processing the reference image.
  • similarity: Calculates similarity scores between the detected faces and reference faces.

• External Functions:

  • Functions like GetProcessCountViaSnapShot, GetUniqueUserCountViaSnapshot, and VmDetection are integrated into the project for advanced runtime features.

Dependencies

This project depends on the following:

• OpenCV: Specifically, a statically compiled version of the OpenCV library is required.
• opencv-rust bindings: Provides the Rust bindings to OpenCV. Check the opencv-rust GitHub repository for more information.

Setup Instructions

Step 1: Compile OpenCV

Before building the project, ensure that OpenCV is compiled statically. Follow the OpenCV documentation for setting up a statically compiled version of the library on your system.

Step 2: Download Models

In the scripts directory, you will find Python scripts that help download offline versions of the MobileNetV2 and CascadeClassifier models. These are required to run the facial detection and comparison components.

Step 3: Build the Project

Once OpenCV is compiled and the models are downloaded, you can build the project using Cargo:

cargo build --release

Step 4: Run the Executable

After building the project, you can run the executable to start the face detection and verification process:

./target/release/T-70.exe

Special Thanks

Special thanks to twistedfall for maintaining the opencv-rust project, which was critical for the development of the T-70 system.

License

T-70 is licensed under the MIT License.

---

Inspired by the Cyberdyne Series 70 Automated Fighting Infantry Unit ("The Terminator").
More information: Cyberdyne Series 70 Terminator