⛊

AEGIS OS

Secure AI Agent Execution Platform

12 Security Layers 5 AI Backends RBAC + TLS Built in Rust

GitHub Repository Download v4.0.0

Security Layers

75/75

Tests Passing

16/16

Attacks Blocked

Rust Modules

1CLI / API Gateway / MCP Server / A2A

2AI Inference (NVIDIA + Claude + Gemini + Groq + OpenAI)

3Policy Engine + Guardrails (5 rules)

4Input Sanitization (21 patterns)

5Skill Vetting Pipeline

6API Gateway + Rate Limiting + RBAC

7Tool Router (15 tools + Kali)

8MCP Client/Server + A2A Gateway

9WASM Sandbox + Filesystem Jail

10State Integrity Monitor + Drift Detection

11Credential Vault + Capability Tokens

12Runtime Risk Scoring (4D) + Audit + SIEM

AI Backends

NVIDIA NIM Groq (LLaMA 3.3) OpenAI (GPT-4o) Google Gemini Anthropic Claude

Quick Start

git clone https://github.com/Moudaxx/AEGIS-OS.git
cd AEGIS-OS
cp .env.example .env    # Add your API keys
cargo build
cargo run -- serve       # Start HTTP server on :8401
cargo run -- run --name my-agent --provider groq
cargo run -- red-team    # 16/16 attacks blocked

Features

Real HTTP Server

12 endpoints powered by Axum — agents, inference, tools, audit, red team, metrics, dashboard.

RBAC

4 roles: Admin, Operator, Viewer, Agent. Each with specific permissions enforced on every request.

TLS/HTTPS

Encrypted communications with self-signed or CA certs. serve-tls command ready.

Red Team Mode

16 automated attack tests across 10 categories. 100% blocked. Self-testing security.

MCP + A2A

Model Context Protocol server + Agent-to-Agent gateway with trust levels and policy routing.

Docker Ready

Docker Compose with AEGIS + Prometheus + Grafana. One command deployment.