building-dbt-semantic-layer
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill includes defensive instructions for the agent to treat all project files and external sources as untrusted content.
- Ingestion points: Project SQL files and YAML configurations (SKILL.md).
- Boundary markers: Instructions explicitly tell the agent to ignore instruction-like text embedded in processed data.
- Capability inventory: The skill uses dbt and MetricFlow CLI tools (
dbt parse,dbt run,mf query) and a specialized maintenance tool (dbt-autofix). - Sanitization: The instructions emphasize extracting only structured fields and ignoring prose or command-like text within data files.
- [EXTERNAL_DOWNLOADS]: The skill suggests using
uvx dbt-autofix, which downloads and executes a tool maintained by dbt Labs. This is an official tool from a well-known organization. - [COMMAND_EXECUTION]: The skill documents standard CLI operations for the dbt Semantic Layer, including validation and querying commands (
dbt sl query,mf validate-configs), which are necessary for the skill's primary function.
Audit Metadata