Developed cross-repository enhancements for LLM error observability, focusing on adobe/spacecat-shared and adobe/spacecat-api-service. Introduced a standardized LLM_ERROR_PAGES audit type, updating data models and unit tests in JavaScript and YAML to ensure consistent error tracking. Defined and registered a new audit results schema, capturing success status, error counts, and detailed error page information, and integrated it into the run-audit workflow. This work established unified audit definitions and reporting across both services, improving operational visibility into LLM failures. Leveraged skills in API development, backend development, and schema definition to deliver features that enable faster triage and standardized monitoring capabilities.