ml-security-tools.md

Recommended Tools for Secure Machine Learning

Threat Modeling AI Systems

• PLOT4ai

ML Source Code Quality & Security

• CodeQL

Third-Party Library Security Vulnerability Scan

• Safety

Secure Jupyter Notebooks

• NB Defense

Benchmark ML Vulnerabilities

• CleverHans

Adversarial Attack & Defense

• Adversarial Robustness Toolbox (ART)
• Advertorch
• Foolbox
• NLP-specific: TextAttack
• Fuzzing: MindArmour

ML Model Vulnerability Scanning

• Armory

LLM Security & Robustness

• Garak - LLM Vulnerability Scanner
• PyRIT - Assess the Robustness of LLM Endpoints

Model & Code Security

• ModelScan - Scan Models for Unsafe Code
• Counterfit - Evaluate Security of AI Models

Privacy & Compliance

• AI Privacy Toolkit

Disclaimer: Thales does not endorse the utilization of any specific tool. The reference to such tools is purely for illustrative and informational purposes, and their use should be exercised with discretion.