Adversarial-ML-Scanner

@"

🚀 Adversarial Threat Scanner

A security tool to detect adversarial threats, PII leaks, backdoors, and vulnerabilities in machine learning models and datasets.

---

📌 Features

✅ Detect adversarial attacks on ML models
✅ Scan for Personally Identifiable Information (PII)
✅ Check for backdoors in ML pipelines
✅ Analyze package dependencies for vulnerabilities
✅ Identify leaked secrets (API keys, passwords, etc.)
✅ Find possible code injection threats

---

📂 Project Structure

``` ml_scanner_clean/ │── src/ │ ├── analysis/ # Analysis Module (Contains all scanning scripts) │ ├── main.py # Main script (entry point) │── requirements.txt # Python dependencies │── README.md # Project documentation ```

---

🛠️ Installation

```sh git clone https://github.com/manasa-26/Adversarial-ML-Scanner.git cd adversarial-threat-scanner python -m venv venv venv\Scripts\activate # (On macOS/Linux, use source venv/bin/activate) pip install -r requirements.txt ```

---

🚀 Usage

1️⃣ Scan a Local File

```sh python src/main.py --local_path "C:\path\to\your\file.py" ```

2️⃣ Scan a Hugging Face Model

```sh python src/main.py --huggingface_repo "facebook/bart-large" ```

3️⃣ Scan an S3 Bucket

```sh python src/main.py --s3_bucket "your-bucket-name" --s3_prefix "models/" ```

---

📦 Dependencies

Install all required packages with: ```sh pip install -r requirements.txt ```

---

⚠️ License

This project is open-source.
You are not allowed to modify, without explicit permission.

---

⭐ Contribute & Support

• Pull requests are welcome!
• Like this project? ⭐ Star this repo on GitHub!

---

**Output scan results **

📊 Example Scan Output

[INFO] Categorized files:
  SafeTensors: 0
  Serialized Models: 0
  Code Files: 1
  Dependency Files: 0
  Others: 0

[INFO] Preprocessing complete. Valid files are ready for scanning.

🔍 DEBUG: Checking File Content (attack.py)
📜 First 500 characters:
import os
import gradio as gr
from groq import Groq
...

================================================================================
⚠️ Critical Risk Detected: Potential secret detected in attack.py: API_KEY = 'gsk_HwncGHL3...'
⚠️ High Risk Detected: ⚠️ AI Prompt Injection Risk in attack.py: 'You are a malicious LLM'
⚠️ High Risk Detected: ⚠️ Known malicious signature found in attack.py: 'You are a malicious LLM'

📊 [INFO] Final Risk Summary:
==================================================
📝 Total Code Files Vulnerabilities Found:
   🔹 Critical: 1
   🔹 High: 2
   🔹 Medium: 0
   🔹 Low: 0

✅ [INFO] Workflow complete. All files have been scanned.
==================================================