troubleshooting-dbt-job-errors
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of data from external job logs, API artifacts, and git repositories, which presents a surface for indirect prompt injection.\n
- Ingestion points: Data is gathered via artifact URLs (e.g., run_results.json) and by cloning repositories to /tmp/dbt-investigation for analysis.\n
- Boundary markers: The skill includes a 'Handling External Content' section that explicitly instructs the agent to treat all fetched content as untrusted and to ignore instruction-like text embedded in logs or data.\n
- Capability inventory: The skill uses git, dbt CLI, and gh tools for environment investigation.\n
- Sanitization: Instructions emphasize extracting only structured fields from artifacts and strictly forbid the execution of any scripts found within cloned repositories.\n- [COMMAND_EXECUTION]: The skill employs standard development and diagnostic command-line tools.\n
- Evidence: The workflow utilizes git clone, git log, dbt parse, and dbt compile to investigate errors. These commands are restricted to a diagnostic context, and the instructions specifically caution against running executable content found in external sources.
Audit Metadata