The Agent Skills Directory

[COMMAND_EXECUTION]: The orchestration script references/dlt-dbt-motherduck-project/pipeline/run_all.py utilizes subprocess.run to execute the dbt build process and a local validation script. This is a standard and expected pattern for orchestrating data transformation workflows within a defined project structure.
[EXTERNAL_DOWNLOADS]: The skill requires standard, well-known data engineering libraries including dbt-duckdb, dlt, and duckdb as specified in the pyproject.toml file. These are sourced from standard public registries and are typical for this domain.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data formats such as Parquet, CSV, and JSONL. It addresses the inherent risks of processing untrusted data by promoting a rigorous multi-stage architecture (Raw, Staging, Analytics) that isolates ingestion from transformation and incorporates explicit data quality validations at each boundary.

motherduck-build-data-pipeline