What problem does it solve? Locating a specific table or dataset in a large AWS data lake is slow and error-prone when assets span Glue Data Catalog, S3 buckets, S3 Tables, and Redshift. This Skill resolves vague references like "the orders table" or an S3 path into concrete catalog entries with database, format, location, and partition keys. ## Core Features & Use Cases - Layered search with early stop: Searches Glue Data Catalog first via search-tables, then S3 reverse lookup, then Redshift, stopping at the first high-confidence match to minimize API calls. - S3 path reverse lookup: Converts an S3 path like s3://bucket/prefix/ into the corresponding Glue table registration. - S3 Tables and Redshift coverage: Enumerates S3 Tables via s3tables list-table-buckets and queries Redshift svv_all_tables for warehouse and lakehouse assets. - Experimental Glue Discovery support: Optionally queries customer-authored catalog context via the Glue Search/GetAsset APIs, with strict untrusted-data handling. - Use Case: A data analyst asks "which table has customer_id?" The Skill searches the Glue catalog, finds three candidate tables, and presents them ranked with format and match reasons. ## Quick Start Ask the assistant to find the table matching a name, keyword, column, or S3 path, for example: find the table containing customer_id in the us-west-2 region.