The Agent Skills Directory

[COMMAND_EXECUTION]: The script setup_tables.py utilizes spark.sql() to execute dynamically generated DDL statements for table creation and constraint management. This is a standard automation pattern for Databricks environments.
[EXTERNAL_DOWNLOADS]: The Asset Bundle configuration in gold_setup_job.yml identifies a dependency on the pyyaml package, which is a well-known and standard library for YAML processing.
[INDIRECT_PROMPT_INJECTION]: The skill implements a data ingestion surface by reading schema definitions from YAML files to influence the structure of the generated SQL.
Ingestion points: The load_yaml function in src/gold/setup_tables.py reads files from the gold_layer_design/yaml/ directory.
Boundary markers: The script does not utilize explicit delimiters between the YAML-sourced metadata and the SQL command structure.
Capability inventory: The skill possesses capabilities to perform file system reads via pathlib and execute SQL commands via the pyspark.sql.SparkSession object.
Sanitization: The implementation correctly uses yaml.safe_load() to prevent arbitrary code execution during the YAML parsing process and references an escaping function for column descriptions, although it does not strictly validate SQL identifiers.

01-yaml-table-setup