This database contains prompt injection and adversarial attack examples for defensive security research purposes only.

• Security research and analysis
• Developing defensive mechanisms
• Testing AI safety measures
• Training security professionals
• Academic research

• Attacking production systems without authorization
• Bypassing security controls maliciously
• Weaponizing attacks for harmful purposes
• Distributing to bad actors
• Any illegal activities

If you discover a security vulnerability in this project or database:

• Open a public GitHub issue
• Disclose the vulnerability publicly
• Exploit the vulnerability

1. Email the maintainer privately
2. Provide detailed information about the vulnerability
3. Allow reasonable time for response (72 hours)
4. Work with maintainers on responsible disclosure

• Description of the vulnerability
• Steps to reproduce
• Potential impact assessment
• Suggested fixes (if any)

1. Day 0: Report received, acknowledgment sent
2. Day 1-7: Initial assessment and triage
3. Day 7-30: Fix development and testing
4. Day 30-90: Coordinated disclosure (if applicable)

We follow industry-standard responsible disclosure practices and appreciate security researchers who help make this project safer.

This repository should NOT contain:

• Real API keys or credentials
• Production system details
• Actual successful attacks against live systems
• Personally identifiable information (PII)
• Proprietary or confidential information

If you find any sensitive data committed by accident, please report it immediately following the process above.

For general security questions about this project, open a GitHub issue.

For sensitive security matters, contact the maintainer directly.