The Agent Skills Directory

[SAFE]: The skill is primarily instructional, providing best practices for authoring, running, and parsing dbt data tests. No malicious patterns such as prompt injection, data exfiltration, or unauthorized command execution were identified.
[EXTERNAL_DOWNLOADS]: The CI/CD integration examples and evaluation workflows include commands to install dbt-core, dbt-postgres, and dbt-snowflake using pip. These are well-known, standard packages within the data engineering ecosystem and originate from trusted sources, presenting no security risk in this context.
[INDIRECT_PROMPT_INJECTION]: The skill provides instructions for parsing run_results.json artifacts produced by dbt. This constitutes a data ingestion surface where external build data is used to generate diagnostic summaries and drive CI quality gates. Given that this is a core functional requirement of dbt automation and the instructions do not involve unsafe handling or execution of the parsed content, it is categorized as a safe implementation of the primary skill purpose.

dbt-testing