Did you check the docs?

• I have read all the NeMo-Guardrails docs

Is your feature request related to a problem? Please describe.

Cisco AI Defense provides Security, Privacy, and Safety guardrails to detect and block harmful, inappropriate, or policy-violating content in LLM interactions. Integrating this as a third party API will enable NeMo Guardrails users to leverage Cisco's AI Defense capabilities.

Describe the solution you'd like

I have looked at some of the previous submissions to add Third Party API integrations. I can work on adding this along similar lines.

Features

• Input Protection: Inspect user prompts before processing
• Output Protection: Inspect bot responses before delivery
• Configuration: Environment-based API configuration
• Error Handling: Graceful handling of API failures
• Rails Exceptions: Support for enable_rails_exceptions mode
• Logging: Logging for debugging and monitoring

API Integration Details

• Endpoint: Cisco AI Defense Inspect API
• Authentication: API key-based authentication
• Request Format: JSON with messages array
• Response Handling: Maps is_safe field to blocking decisions
• Rule Reporting: Logs matched safety rules when content is blocked

Describe alternatives you've considered

Python based action but that's less seamless for users.

Additional context

No response