The Agent Skills Directory

[SAFE]: The skill functions as an educational and operational guide for data engineers. The instructions and code examples provided are consistent with industry-standard practices for infrastructure, logic, and data quality debugging.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill contains Python code for querying metadata and lineage APIs (Marquez, DataHub). These operations use the requests library to fetch dependency graphs from internal services, which is expected behavior for its primary purpose of blast radius assessment. No unauthorized data transmission to external or untrusted domains was found.
[COMMAND_EXECUTION]: Several sections provide examples of using CLI tools such as dbt, kubectl, and airflow. Additionally, Python snippets include subprocess.run to execute dbt tests. These are standard administrative and debugging commands used in the context of pipeline investigation.
[INDIRECT_PROMPT_INJECTION]: The skill describes methods for parsing and analyzing system logs (Airflow, Spark, application logs). While reading external data can present an injection surface, the provided scripts perform specific extraction and filtering tasks (e.g., using jq or pandas) rather than open-ended instruction processing, minimizing risk.

de-root-cause-analysis