If you discover a security vulnerability in these tools, please report it privately.

Email: scott@perfecxion.ai

Please include:

• Description of the vulnerability
• Steps to reproduce
• Potential impact
• Suggested fixes

• Initial Response: Within 48 hours
• Assessment: Within 7 days
• Resolution: Based on severity

These tools are designed for authorized security testing and analysis of ML models you own or have permission to test.

✅ Permitted:

• Security testing of your own ML models
• Authorized penetration testing with permission
• Security research in controlled environments
• Educational purposes
• Defensive security analysis

❌ Not Permitted:

• Unauthorized testing of production ML systems
• Attacking third-party ML services without permission
• Any activity violating terms of service
• Malicious use or exploitation

• Tests adversarial robustness of models
• Use only on models you own or have permission to test

• Analyzes model architecture and parameters
• May reveal sensitive model details - handle output securely

• Detects potential poisoning in training data
• Run only on datasets you're authorized to analyze

• Validates TensorRT export security
• Ensures safe model deployment

• Analyzes PyTorch checkpoint files
• May contain sensitive information - handle securely

• Audits Triton Inference Server configurations
• Use only on infrastructure you own or manage

• Email: scott@perfecxion.ai
• Alternative: scthornton@gmail.com