Runs Athena SQL queries across Glue, S3 Tables, and Redshift catalogs with proper workgroup selection and cost tracking. It classifies statements before execution (SELECT versus DROP), delegates fuzzy table lookups to the finding skill, and profiles unfamiliar tables before analytical queries. The standout bit is how it handles catalog-specific addressing quirks (Redshift lowercase identifiers, cross-catalog qualifiers) and enforces confirmation gates for destructive operations. Works through AWS MCP server when available, falls back to CLI otherwise. Most useful when you need to query federated sources without manually wrangling workgroup configs or accidentally running unscoped queries that scan entire tables.
npx skills add https://github.com/aws/agent-toolkit-for-aws --skill querying-data-lake