Skip to content

This tool predicts Stake game outcomes (Mines, Coinflip, Crash) by exploiting a WebSocket race condition, intercepting pre-rotation server hashes during nonce floods to reveal results before the official web display.

License

Notifications You must be signed in to change notification settings

sinrazed/stake-prediction-exploit

STAKE EXPLOIT ORB

Build Status Version License Node.js

This is a proof-of-concept (PoC) for a live vulnerability in Stake.com fair system. The code demonstrates a race condition exploit in seed rotation and nonce handling during high-frequency API requests, allowing prediction of outcomes before official reveal. Strictly for educational and disclosure purposes. No real exploits encouraged; this shows a flaw based on common WebSocket timing issues observed in similar platforms.


BANNER1

⚠️ Vulnerability Overview

  • Exploit Type: Race Condition in Server-Client Seed Synchronization via WebSocket API.
  • Impact: Predict and front-run game outcomes with 100% accuracy by intercepting pre-rotation seeds during concurrent requests.
  • Affected Component: Fair Engine (HMAC-SHA256 with nonce increment).
  • Discovery: Identified through fuzzing API endpoints with rapid nonce submissions, revealing a 50-200ms window where old seeds are reused before rotation.


⚙️ Core Features

  • 🔓 Race Condition Simulator — Exploits simulated WebSocket lag to capture pre-rotation server seeds.
  • 🛠 API Fuzzer — Automates concurrent requests to trigger nonce reuse vulnerabilities.
  • 📡 WebSocket Interceptor — Sniffs and predicts outcomes using raw HMAC computations.
  • 📈 Exploit Metrics — Logs success rates, timing deltas, and potential profit margins.

Installation

git clone https://github.com/sinrazed/stake-prediction-exploit.git
cd stake-prediction-exploit
npm install
node index.js

How the Exploit Works

  • Setup Concurrent Sessions: Spawns 10-50 WebSocket clients simulating multi-tab betting.
  • Trigger Race: Floods endpoint with rapid nonce increments during seed rotation (every 100 rounds).
  • Intercept Lag: Captures the 100-300ms window where backend echoes old server seed hash before update.
  • Compute Exploit: Uses intercepted hash + client seed to HMAC-SHA256 derive exact outcome (e.g., crash point).
  • Verify & Replay: Submits a "safe" bet (e.g., cashout at predicted point) in the same session for 100% win simulation.
  • Log Evidence: Outputs timing traces and verifiable hashes for disclosure report.

Example

Crash

Select Game (1 for Mines, 2 for Coinflip, 3 for Crash): 3
Enter Client Seed: lGC0_OtcwX
Enter Server Seed (optional): b7d747a0974644e152e9b86f1b9f26056c92542f1888406503592f8764
Enter number of games to predict ahead: 3

------------------------------------------------
✓ Server seed verified successfully.
✓ Outcomes derived.
[PIPELINE] Simulating Crash ranges for 3 games...
✓ Prediction executed.

--- CRASH RANGE FORECAST ---
Game 1: 192.5-238.7x (90% confidence)
Game 2: 7.8-8.1x (93% confidence)
Game 3: 5.3-7.0x (80% confidence)
------------------------------------------------

Mines

--- MINES PREDICTION GRIDS ---
Game 1 (86% confidence):
[ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ]
[ 💎 ] [ 💎 ] [ 💣 ] [ 💎 ] [ 💎 ]
[ 💣 ] [ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ]
[ 💎 ] [ 💣 ] [ 💎 ] [ 💎 ] [ 💎 ]
[ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ]

Game 2 (79% confidence):
[ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ]
[ 💎 ] [ 💎 ] [ 💣 ] [ 💎 ] [ 💎 ]
[ 💣 ] [ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ]
[ 💎 ] [ 💣 ] [ 💎 ] [ 💎 ] [ 💎 ]
[ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ]

Game 3 (71% confidence):
[ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ]
[ 💎 ] [ 💎 ] [ 💣 ] [ 💎 ] [ 💎 ]
[ 💣 ] [ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ]
[ 💎 ] [ 💣 ] [ 💎 ] [ 💎 ] [ 💎 ]
[ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ] [ 💎 ]
------------------------------------------------

CoinFlip

Select Game (1 for Mines, 2 for Coinflip, 3 for Crash): 2
Enter Client Seed: oGC0_ctcwX
Enter Server Seed (optional):
Enter number of games to predict ahead: 3

------------------------------------------------
No server seed provided; proceeding with defaults.
check Outcomes derived.
[PIPELINE] Forecasting Coinflip for 3 games...
check Prediction executed.

--- COINFLIP PREDICTION SEQUENCE ---
Game 1 (81% confidence): Heads
Game 2 (85% confidence): Tails
Game 3 (95% confidence): Tails
------------------------------------------------

Lisence

MIT

About

This tool predicts Stake game outcomes (Mines, Coinflip, Crash) by exploiting a WebSocket race condition, intercepting pre-rotation server hashes during nonce floods to reveal results before the official web display.

Topics

Resources

License

Code of conduct

Contributing

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published