managing-data-catalogs

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides configuration patterns and architectural guides for data catalogs using well-known services and industry-standard libraries. No malicious code, obfuscation, or unauthorized access patterns were identified.- [PROMPT_INJECTION]: The skill facilitates the ingestion of data from external catalogs and cloud storage, which represents an indirect prompt injection surface (Category 8).
  • Ingestion points: Multiple files including duckdb-catalog.md, rest-catalog.md, and hive-metastore.md describe fetching and processing table data.
  • Boundary markers: Not explicitly present in the documentation snippets.
  • Capability inventory: The skill enables SQL execution via DuckDB and Python-based data processing via Pandas and PyIceberg.
  • Sanitization: As these are connectivity and setup examples, input validation and data sanitization are not explicitly addressed.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 06:13 PM