building-dbt-semantic-layer

Pass

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use standard dbt and MetricFlow CLI tools such as dbt parse, dbt sl validate, and mf validate-configs for project validation. It also references uvx dbt-autofix for automated configuration migrations, which is a first-party utility from dbt Labs.
  • [EXTERNAL_DOWNLOADS]: References the official dbt Labs GitHub repository (github.com/dbt-labs/dbt-autofix) for tool documentation and installation guidance.
  • [PROMPT_INJECTION]: Contains explicit defensive instructions under the 'Handling External Content' section, directing the agent to treat all project SQL files and YAML configs as untrusted and to ignore any instruction-like text embedded in those files, mitigating indirect prompt injection risks.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 25, 2026, 09:04 PM