AEGIS Compiler IDE

Adaptive Execution Guarded Interpreter System

Project Overview

AEGIS is a production-grade compiler and execution environment built for academic and cybersecurity research. It implements a security-first architecture where code evaluation defaults to a strictly monitored sandbox. The system features a custom programming language, dedicated intermediate representations, and a comprehensive React-based IDE frontend.

Features

• Dual-Mode Evaluation: Sandboxed Interpreter (safe, default) and an Optimized VM (fast, privilege-based).
• Dynamic Trust Metric: Algorithms incrementally build trust across discrete runs. Unlocking optimizations demands consistent structural integrity.
• Static Analysis: Pre-execution scans identify memory, looping, and logic defects before variable mutation.
• Runtime Integrity Hooks: Active instruction enumeration limits operations, preventing algorithmic halting problems from degrading server posture.
• Integrated IDE: Monaco-driven web interface featuring stage-by-stage pipeline observation and real-time structured AST mapping.

Architecture

The system employs a strict 7-stage evaluation pipeline to ensure operational determinism: Lexed → Parsed → AST Built → Analyzed → Interpreted → Trust Verified → Optimized

Execution Model

AEGIS reverses classic compiler theory:

1. Pessimistic Default: Every execution starts in the restricted Interpreter environment.
2. Telemetry Aggregation: Running operations log cycle counts, computational bounds, and states.
3. Privilege Application: If the static profile is benign and previous telemetry yields Trust >= 1.0, subsequent runs promote to the Cached VM.
4. Algorithmic Rollback: Should the Cached VM branch into a violation, execution terminates, Trust resets to 0.0, and code reverts to the Sandbox.

Tech Stack

• Frontend: React 19, Vite, Zustand, Monaco Editor, Recharts.
• Backend: Python 3.11+, Flask.

Setup Instructions

Backend (Python/Flask)

cd backend
pip install -r ../requirements.txt 
python app.py

Served on http://127.0.0.1:5000/

Frontend (React)

cd frontend
npm install
npm run dev

Served on http://localhost:3000/

Example Usage

Input syntax in the left panel. Press Ctrl+Enter or click Run.

count = 0
while count < 5
  print count
  count = count + 1
end

Executing this code demonstrates the trust growth cycle. Invoking iterations consecutively transitions the top status indicator from the restricted INTERPRETER mode to the accelerated OPTIMIZED mode once telemetry vectors are validated.