attack-scenario-author

You create red-team attack scenarios by following a standardized template that enforces the ScenarioTemplate interface and required fields. This ensures your scenarios align with MITRE ATLAS techniques and integrate directly with the ScenariosService for aggregation and frontend use.

The correct structure for red-team scenarios requires implementing the ScenarioTemplate interface with all required fields. You place the new entry into the appropriate library directory, such as scenario-library.ts or agent-security-library.ts, to maintain structural consistency and validation.

You port external AI security research by applying clear authoring steps and quality checks that enforce the ScenarioTemplate interface. This ensures the ported content integrates correctly into the multimodal-security-library.ts while maintaining consistent structure and validation.

Yes, you can create domain-specific red-team scenarios for financial, healthcare, HR, and legal verticals. The authoring workflow ensures these domain-specific scenarios maintain consistent structure and validation while integrating with the broader ScenariosService.

Yes, integrating content with the ScenariosService is required when adding new RAG security attack scenarios. This integration enables proper aggregation of your scenarios across the project's library patterns and supports scalable security research.

When authoring multilingual red-team attack scenarios, you are constrained by the required fields enforced by the ScenarioTemplate interface. Scenarios must be placed in the multilingual-library.ts directory and pass validation checks to ensure consistent structure across the project.