V8 Randomness Predictor

A Python-based predictor using z3 to reverse-engineer and predict random values generated by Math.random in V8 with 100% accuracy. This script leverages a series of Math.random() values (min 5) to deduce future states.

💬 Discord · 📜 ChangeLog · ⚠️ Report Bug · 💡 Request Feature

---

⚙️ Installation

• Requires: Python 3.8+
• Set up a virtual environment:

  python3 -m venv venv

• Activate the environment:

  # Windows
  venv\Scripts\activate
  # macOS/Linux
  source venv/bin/activate

• Install requirements:

  pip install -r requirements.txt

• Install nodejs (optional)

---

🔥 Features

• Debug Logging: Debug logging through logmagix that tracks each prediction stage.
• Z3 Integration: Uses Z3 to reverse the XORSHIFT+128 algorithm behind Math.random in V8.
• Custom Sequence Input: Optionally, input your own sequence of Math.random() values for testing.
• Automated Prediction: Computes the next Math.random() output based on previous values.
• Error Handling: Verifies the Z3 model's validity and provides detailed feedback.

---

🔍 Usage

1. Generate Random Numbers: Use V8 (Node.js, Chrome, etc.) to produce a sequence of random values with Math.random(). An example sequence can be generated by running:

   Array.from(Array(5), Math.random);

   Optionally, set a seed in Node.js to reproduce the same random sequence:

   node --random_seed=1337
   Array.from(Array(5), Math.random)

2. Run the Predictor: Input the random sequence into the script. For example:

   python main.py --sequence "[0.9311600617849973, 0.3551442693830502, 0.7923158995678377, 0.787777942408997, 0.376372264303491]"

   Or for testing purposes you can run the python code directly (requires NodeJS):

   python main.py

3. Output Example:

   Using Sequence: [0.376372264303491, 0.787777942408997, 0.7923158995678377, 0.3551442693830502, 0.9311600617849973]
   States: {'se_state0': 4268050313212552111, 'se_state1': 6942842836049070467}
   Successfully guessed result: 0.23137147109312428
   

---

📹 Demo

---

📜 ChangeLog

v1.0.0 ⋮ 11/10/2024
! Initial release with Math.random predictor

---

Inspired by: PwnFunction