⚠️ DEPRECATED – Old Prototype

This repository contains the first experimental version of Sentinel AI. It has now been fully replaced by the modern, stable Sentinel AI v2.

Please use the updated version here: 👉 https://github.com/DarekDGB/Sentinel-AI-v2

Sentinel AI v2 includes:

• improved risk detection
• cleaner architecture
• full test suite
• better integration with DQSN and ADN v2
• future-proof design for the 5-Layer Quantum Shield Network

This repo remains available only for historical reference.

DigiByte-Sentinel-AI

AI-based anomaly detection & threat-monitoring system for the DigiByte blockchain

⸻

Overview

DigiByte-Sentinel-AI is a modular, lightweight, real-time anomaly detection engine designed to enhance the security of the DigiByte blockchain. It performs continuous analysis of transactions, block patterns, signature entropy, and network behavior using a hybrid statistical–ML model.

The goal is to provide the DigiByte ecosystem with an autonomous early-warning system against: • irregular transaction bursts • potential 51% attack indicators • address behavior drift • low-entropy or weak digital signatures • chain manipulation attempts • suspicious miner activity • cross-chain exploit patterns

⸻

Key Features

1. Real-time anomaly scoring

Extracts 6 core metrics per transaction and assigns a security score from 0.0 to 1.0.

2. Intelligent risk classification • normal • elevated • high • critical

3. Lightweight hybrid model

No GPU required — optimized for running on nodes, servers, or monitoring dashboards.

4. REST API endpoint

Enables wallets, explorers, miners, and exchanges to query the engine: POST /sentinel/analyze 5. Event logging

Every analyzed transaction includes a timestamp and structured JSON log output.

⸻

Architecture

Data ingestion • Pulls transaction metadata • Computes statistical and behavioral metrics • Feeds vectorized data into the scoring model

Scoring engine

Uses blended weights across: • amount deviation • address recurrence • low-entropy signature detection • frequency anomaly • miner pattern recognition • multi-factor irregularities

Output

JSON object containing: • anomaly score • classification • timestamp • input metrics Installation 1. Clone repository 2. Install dependencies: pip install fastapi uvicorn numpy 3. Run the API: uvicorn sentinel_ai:app --host 0.0.0.0 --port 8000 API Example

Request: { "txid": "abc123", "amount": 42000, "inputs": 3, "outputs": 1, "signature_entropy": 0.14, "address_reuse": 1 } Response: { "txid": "abc123", "anomaly_score": 0.78, "classification": "high", "metrics": { "amount_zscore": 2.13, "freq_score": 0.44, "entropy_score": 0.86 } } Security Goals • Strengthen DigiByte’s defense against evolving smart-attack patterns • Provide transparent, open-source monitoring tools for the ecosystem • Offer exchanges and explorers automated anomaly detection • Support future upgrades toward quantum-resilient threat analysis

⸻

Status

This is an open development prototype and can be extended by DigiByte core developers, security researchers, and the wider community.

⸻

License

MIT License — free to use, modify, integrate, or expand.

⸻

Maintainer

Created by DarekDGB Visionary security concept contributor for the DigiByte ecosystem.