Project Intelligence Vault

Applied AI/ML + Cybersecurity Systems

A curated archive of production-oriented projects spanning malware analysis, secure communication, DevSecOps automation, steganography, process anomaly detection, and environmental ML forecasting.

Byte-Brain

PythonScikit-learnPEfile

Byte-Brain is an offline static PE malware scanner built for high-confidence risk assessment without cloud dependency.

It uses a Random Forest classifier trained on EMBER 2018 features to estimate probability-based maliciousness scores.

The pipeline emphasizes explainable predictions so analysts can understand which static indicators raised risk.

Key Features

  • Offline PE scanning for privacy-preserving malware triage
  • Probability-based maliciousness scoring using EMBER-style features
  • Explainable outputs highlighting high-impact static signals
View GitHubDeep Dive & Demo

Secure Secrets

ReactNode.jsWeb Crypto APIExpress.js

Secure Secrets is a zero-knowledge, end-to-end encrypted secret sharing platform focused on privacy-first communication.

All encryption happens client-side using AES-GCM 256 with the Web Crypto API before any network transmission occurs.

The platform also supports burn-on-read flows and EXIF metadata stripping for safer media and secret exchange.

Key Features

  • Client-side AES-GCM 256 encryption before transport
  • Burn-on-Read links for one-time secret retrieval
  • EXIF metadata stripping to reduce accidental data leaks
View GitHubDeep Dive & Demo

Git-Sentinel

PythonRegexDevSecOps

Git-Sentinel is a lightweight pre-commit security scanner designed to stop secret leaks before they enter version history.

It hooks into the commit flow, inspects staged diffs, and flags risky patterns using regex-based detections.

By failing unsafe commits early, it improves baseline DevSecOps hygiene across personal and team repositories.

Key Features

  • Interception of commit workflow before history is written
  • Regex signatures for API keys, tokens, and secret patterns
  • Actionable alerts that block unsafe commits by default
View GitHubDeep Dive & Demo

Shadow-Pixel

PythonCryptographySteganographyAES-256-GCM

Shadow-Pixel is a CLI utility for invisible cryptographic steganography in RGB images.

It encrypts secrets with AES-256-GCM and hides ciphertext inside Least Significant Bits (LSB) to preserve visual fidelity.

The entire workflow runs offline, enabling controlled covert communication and secure local experimentation.

Key Features

  • AES-256-GCM encryption before steganographic embedding
  • LSB-based data hiding with minimal visual distortion
  • Offline CLI workflow for secure local operation
View GitHubDeep Dive & Demo

Aura Process Guardian

PythonpsutilStatisticsZ-Score Modeling

Aura Process Guardian is an unsupervised OS-level anomaly detector for real-time process monitoring.

It learns baseline behavior from runtime telemetry and flags sustained deviations using statistical Z-score modeling.

The system is built for early detection of suspicious process activity without requiring labeled attack datasets.

Key Features

  • Real-time process telemetry collection from live systems
  • Adaptive baseline learning for normal behavior profiles
  • Sustained anomaly detection via Z-score thresholds
View GitHubDeep Dive & Demo

Flood Prediction ML Model

PythonScikit-learnPandasXGBoost

Flood Prediction ML Model estimates flood probability from environmental and hydrological factors.

The project applies feature engineering and compares model behavior across Random Forest and XGBoost regressors.

Its objective is to support risk-aware planning by turning raw variables into interpretable predictive outputs.

Key Features

  • Environmental feature engineering for stronger signal extraction
  • Model comparison between Random Forest and XGBoost
  • Probability-oriented outputs for flood risk estimation
View GitHubDetails