Skip to content

CHANGELOG.md

Latest commit

 

History

History
98 lines (68 loc) · 3.1 KB

CHANGELOG.md

File metadata and controls

98 lines (68 loc) · 3.1 KB

🧾 CHANGELOG

Project: Super RNN — Advanced AI System for Kali Linux Security Analysis
Author: GUIAR OQBA
Initial Release: 2025
Version: v1.0.0 (Stable Build)

🧩 Version 1.0.0 — Initial Release (2025-10-14)

🚀 Overview

The first stable release of Super RNN, an advanced artificial intelligence engine written in pure C for security analysis on Kali Linux systems.
This release introduces the core RNN architecture, Adam optimizer, and AI-powered system threat analysis.

🧠 Core Features Implemented

  • Custom Recurrent Neural Network (RNN) implemented entirely from scratch in C
  • Forward & Backward Propagation (BPTT) with Mean Squared Error loss
  • Adam Optimizer with momentum, velocity, and adaptive learning rates
  • Xavier Weight Initialization for stable gradient flow
  • Early Stopping to prevent overfitting
  • Dropout Regularization for improved generalization
  • Data Normalization & Denormalization for consistent training

🔐 Security & System Analysis

  • Configuration File Analysis: /etc/passwd, /etc/group, etc.
  • Log Analysis: /var/log/syslog, /var/log/auth.log, etc.
  • Command Learning: Learns common Linux utilities
  • Threat Prediction: Estimates system threat level (LOW–MEDIUM–HIGH)
  • Automated Reporting: Generates detailed reports in /tmp/security_report.txt

⚙️ System Integration

  • System Data Collection: From /etc, /proc, /var/log
  • Real-Time Monitoring: CPU, memory, and disk usage
  • Command Execution: Learns and executes Linux commands for analysis
  • Report Generation: Creates structured security summaries

🧰 Build & Execution

  • Compatible with Linux-based systems (tested on Kali Linux)
  • Compiles using gcc with optimization flags: 
    gcc -O3 -Wall -Wextra -o super_rnn super_rnn.c -lm


Executable modes :

  ```bash
./super_rnn
sudo ./super_rnn

📄 Output Example

Displays training progress (loss values per epoch)

Shows predicted vs. target outputs

Reports system security level

Confirms saved report path

🧾 Dependencies

Standard C libraries (math.h, stdlib.h, stdio.h, etc.)

Linux system files (/proc, /etc, /var/log)

📦 Release Notes Summary Component Status Description 🧠 Neural Engine ✅ Implemented Custom RNN core from scratch ⚙️ Optimizer ✅ Implemented Adam Optimizer with bias correction 🔐 Security Module ✅ Implemented System threat prediction & analysis 📊 Reporting ✅ Implemented Auto report generation 🧩 Linux Integration ✅ Implemented File, log, and command learning 🧮 Data Processing ✅ Implemented Normalization, clipping, encoding 🧰 Compilation ✅ Ready GCC build tested on Kali Linux 📈 Version 🟢 Stable Initial public release 👤 Author & License

Author: GUIAR OQBA

Email: techokba@gmail.com

License: MIT License

Made with 💻 & ❤️ from Algeria 🇩🇿

🧠 “Super RNN marks the beginning of AI-driven Linux security — built from scratch, for hackers, analysts, and researchers.”