Apppublic

aditisingh4/cyber_security_env

sourceHugging Faceupdated 6mo agoView on Hugging Face

0likes

App README

Cyber Security Defense OpenEnv Environment

Overview

This project is a real-world OpenEnv environment that simulates a cyber security defense system where an AI agent must protect a server from cyber attacks such as malware, DDoS attacks, and suspicious IP activity.

The agent interacts with the environment using actions and receives rewards based on how well it protects the system.

Environment State (Observation Space)

The agent receives the following state:

—system_health (0–100)
—cpu_usage
—network_traffic
—suspicious_ips
—malware_detected (True/False)
—ddos_attack (True/False)
—time_step

Actions (Action Space)

The agent can take the following actions:

—monitor
—block_ip
—scan_file
—quarantine_file
—enable_firewall
—shutdown_server

Reward Design

The agent receives rewards based on performance:

Action	Reward
Block suspicious IP	+1
Detect malware	+2
Remove malware	+3
Stop DDoS	+2
Wrong action	-0.5
Unnecessary shutdown	-2
System damage	Negative

Tasks

Easy Task

Detect and remove malware.

Medium Task

Stop DDoS attack.

Hard Task

Protect the system for 50 steps.

Grading

Each task is graded from 0.0 to 1.0 based on performance.

How to Run

bash

docker build -t cyber-env .
docker run cyber-env