Skip to content

Champbreed/SENTINEL-RAG

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Sentinel-RAG 🛡️💡

A Grounded Knowledge Assistant for Auditable Kernel Security & Neurodiverse Accessibility.

Sentinel-RAG License: MIT

Sentinel-RAG is a specialized accessibility tool designed to reduce cognitive load for individuals—including those with ADHD, autism, and dyslexia—interacting with complex Linux Kernel telemetry. Built for the Microsoft Neurodiversity Challenge, it leverages Python-driven Machine Learning to transform dense, high-anxiety technical data into structured, calm, and actionable insights.

🧠 Features for Neurodiversity

  • Cognitive Load Reduction: Uses Python logic to parse raw eBPF logs into a structured, distraction-free TUI (Terminal User Interface).
  • MLT-Driven Cognitive Compression: Our Python backend (rai_audit.py) uses Machine Learning Text (MLT) logic to dynamically adjust information density. This prevents "Analysis Paralysis" by allowing users to choose the depth of information that matches their current cognitive energy.
  • Level 0 (Technical): Full breakdown of 7+ security targets (such as: capabilities.7, bpf.2).
  • Level 2 (Summarized): Intelligent compression into a single "SYSTEM SUMMARY" to prevent information overload.
  • Grounded Remediation Hints: Provides non-anxiety-inducing, step-by-step tasks to fix non-compliance issues.
  • Visual Focus Zones: Dedicated spatial areas for Vision, Kernel Monitoring, and AI Coaching to prevent sensory overwhelm.

🛠️ Technical Stack & Focus

  • Python (Core Intelligence): The heart of the system. Uses Python for the Responsible AI (RAI) backend to analyze kernel data and generate grounded responses.
  • Microsoft RAI Toolbox: Integrated to ensure all AI outputs are safe, ethical, and supportive.
  • Go (Golang): Provides the high-performance, accessible interface shell using Bubble Tea & Lipgloss.
  • RAG Integration: Security hints are strictly grounded in authoritative kernel "Rulebooks" (capabilities.7, bpf.2) to eliminate AI hallucinations.

🚀 Getting Started

Follow these steps to set up the Sentinel-RAG environment on your local machine.

📋 Prerequisites

  • Python 3.10+
  • Go 1.21+
  • Linux/WSL2 Environment (Required for Kernel telemetry access)

⚙️ Installation & Setup

  1. Clone the Repository
git clone [https://github.com/champbreed/sentinel-rag.git](https://github.com/champbreed/sentinel-rag.git)
cd sentinel-rag
  1. Configure Python Environment
  • Initialize the virtual environment to isolate Machine Learning dependencies.
python3 -m venv venv
source venv/bin/activate
pip install --upgrade pip
pip install pandas responsibleai
  • Inspect the RAI audit logic and groundness scoring (use nano, vim, or cat)
nano scripts/rai_audit.py
  1. Verify MLT Compression Engine
  • View detailed technical audit (Level 0)
python3 scripts/rai_audit.py --level 0
  • View AI-compressed system summary (Level 2)
python3 scripts/rai_audit.py --level 2
  1. Launch the Accessibility TUI
  • Start the interactive Go-based interface.
go run main.go

🛡️ Safety & Ethics

Sentinel-RAG adheres to Microsoft's Responsible AI principles. It employs a Strict-Groundedness policy, ensuring the AI Coach only suggests security remediations verified against official Linux Kernel documentation, preventing harmful or misleading "hallucinations" in critical security contexts.

​Privacy-First Design: Sentinel-RAG processes kernel telemetry locally via eBPF. Sensitive system data remains within the local environment, ensuring that security audits do not compromise system privacy.

About

Sentinel-RAG: A Python & Go-powered accessibility assistant for Linux Kernel auditing. Reduces cognitive load through MLT-driven compression and grounded, non-anxiety-inducing remediation hints.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors