The Agent Skills Directory

[SAFE]: The skill defines a legitimate persona for a Data Engineer, focusing on data pipeline reliability, observability, and architecture (Medallion Architecture).
[SAFE]: The provided code snippets for PySpark, Delta Lake, and Great Expectations are standard implementations for data engineering tasks and do not contain malicious commands.
[COMMAND_EXECUTION]: The skill specifies the use of standard CLI tools including bash and git to perform engineering tasks, which is consistent with its stated purpose.
[INDIRECT_PROMPT_INJECTION]: The agent is designed to ingest external data from JSON files and Kafka streams. While this provides a theoretical surface for indirect prompt injection, the risk is inherent to the role's data-processing nature.
Ingestion points: ingest_bronze reads from JSON files; stream_bronze_orders reads from Kafka topics.
Boundary markers: None identified.
Capability inventory: The skill has access to bash, git, and file system operations.
Sanitization: Standard data schema enforcement is suggested, but no specific prompt-injection sanitization is present.

data-eng