Vedi23/internet-diagnosis-env

Your README.md should now start like this:

title: Internet Diagnosis Env emoji: 🌐 colorFrom: blue colorTo: green sdk: docker pinned: false license: mit ---

🌐 Internet Outage Diagnosis Environment

An OpenEnv environment where an AI agent diagnoses real-world internet connectivity failures.

What is this?

This environment simulates a real-world network support scenario. An AI agent receives network diagnostic data (ping times, packet loss, signal strength, DNS status) and must identify the root cause of internet outages.

This is a task humans actually do every day — network engineers, ISP support staff, and IT helpdesks diagnose these issues constantly.

Observation Space

What the AI sees at each step:

Action Space

What the AI can do:

Along with:

• —failing_component — what specifically is broken
• —suggested_fix — how to fix it
• —confidence — how sure the AI is (0.0 to 1.0)

Tasks

Task 1 — Easy: Single Point Failure

• —2 scenarios with one obvious failure
• —Max 3 steps per episode
• —Pass score: 0.6

Task 2 — Medium: Multi-Layer Diagnosis

• —2 scenarios requiring careful analysis
• —Max 5 steps per episode
• —Pass score: 0.5

Task 3 — Hard: Complex Intermittent Failure

• —2 scenarios with overlapping issues
• —Max 7 steps per episode
• —Pass score: 0.4

Reward Function

Baseline Scores

Running baseline.py with llama-3.1-8b-instant via Groq:

Setup Instructions

Option 1 — Run Locally

Step 1 — Install dependencies:

pip install -r requirements.txt

Step 2 — Start the server:

python server.py

Step 3 — Run baseline:

export GROQ_API_KEY=your-key-here
python baseline.py

Step 4 — Visit interactive docs:

http://127.0.0.1:7860/docs

Option 2 — Run with Docker

Step 1 — Build:

docker build -t internet-diagnosis-env .

Step 2 — Run:

docker run -p 7860:7860 internet-diagnosis-env

API Endpoints

Project Structure

internet-diagnosis-env/
├── models.py        # Pydantic data models
├── scenarios.py     # Network failure scenarios
├── tasks.py         # Tasks and graders
├── environment.py   # Core step/reset/state logic
├── server.py        # FastAPI server
├── baseline.py      # Baseline inference script
├── openenv.yaml     # OpenEnv metadata
├── Dockerfile       # Container setup
├── requirements.txt # Dependencies
└── README.md        # This file