What problem does it solve? AWS Clean Rooms collaborations fail with opaque access denied errors and missing CloudWatch logs for custom ML training and inference jobs, and tracing the root cause across IAM roles, S3 bucket policies, KMS keys, and Lake Formation permissions is time-consuming and error-prone. ## Core Features & Use Cases - Permission Debugging: Systematically checks IAM role policies (inline and managed), S3 bucket policies, KMS key policies, Lake Formation grants, and cross-account trust configurations to pinpoint access failures. - ML Logging Diagnostics: Verifies Configured Model Algorithm Association privacy configurations, ML Configuration roles, and CloudWatch log groups to explain why training or inference job logs are not published. - Structured Diagnosis Reports: Produces a standardized output with error classification, pass/fail diagnostic results, root cause, and exact CLI fix commands. - Use Case: A customer reports that their Clean Rooms ML trained model shows PUBLISH_FAILED for logs. The Skill walks through the privacy configuration, ML Configuration role permissions, and log group checks, then identifies the missing logs:PutLogEvents permission and provides the exact IAM policy fix. ## Quick Start Diagnose why my Clean Rooms membership is getting an access denied error on a protected query in region us-east-1 with membership ID abc123.