answering-natural-language-questions-with-dbt

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFE

Full Analysis

[SAFE]: The skill instructs the agent on handling natural language data queries by utilizing dbt's Semantic Layer, existing metrics, or custom SQL. It implements proper boundaries by requiring the agent to use tool-specific commands or locally filtered jq commands on manifest.json and catalog.json instead of reading large files directly into context. No external code execution, data exfiltration vectors, prompt injection, or unsafe credential patterns are present.

Audit Metadata

Risk Level

SAFE

Analyzed

Sep 14, 2026, 08:28 PM