The Agent Skills Directory

[SAFE]: The skill consists of documentation and standard SQL examples for Snowflake features such as Dynamic Tables, Streams, and Snowpipe. All analyzed patterns, including command execution and data handling, are consistent with legitimate data engineering workflows and the stated purpose of the skill.- [INDIRECT_PROMPT_INJECTION]: The skill documents data pipelines that ingest external data, which is a standard and necessary feature for Snowflake data engineering.
Ingestion points: Data is ingested from external stages via Snowpipe and Snowpipe Streaming into raw_events tables as described in the workflow sections of SKILL.md.
Boundary markers: The instructions recommend explicit casting of variant fields (e.g., ::STRING) and using STRIP_NULL_VALUES to manage data consistency and boundaries during ingestion.
Capability inventory: The skill defines workflows using Snowflake SQL DDL/DML, Dynamic Tables, and Tasks to transform and process ingested data.
Sanitization: Data is processed using standard relational SQL types and structure; the skill focuses on architectural best practices rather than specific sanitization of natural language instructions within the data content.

snowflake-data-engineering