review-dashboard-widgets
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data during its audit process, which could theoretically contain adversarial content.
- Ingestion points: The agent is instructed to read
dashboard.json(generated from Databricks CLI) and view definitions fromSHOW CREATE TABLEcalls. - Boundary markers: Absent. The instructions do not specify using delimiters to isolate these inputs.
- Capability inventory: The skill uses the
databricksCLI for data retrieval anddatabricks-sqlsource blocks for query execution. - Sanitization: Absent. The skill reconstructs queries directly from parsed metadata.
- [DYNAMIC_EXECUTION]: The skill generates SQL code at runtime based on the dashboard configuration and then executes it to verify results.
- Evidence: Step 4 involves reconstructing a query from dataset fields, filters, and encoding, then running it in a
databricks-sqlblock to record the results.
Audit Metadata