motherduck-load-data
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of data from external formats and remote locations, creating a surface where malicious instructions embedded in the source data could potentially influence the agent's behavior.
- Ingestion points: Usage of
read_csv,read_parquet,read_json,delta_scan,iceberg_scan, andATTACHfor remote databases (found inSKILL.mdandreferences/INGESTION_PATTERNS.md). - Boundary markers: Instructions do not include specific boundary markers or warnings to disregard instructions within the ingested data.
- Capability inventory: The skill can execute SQL for table creation, data insertion, and credential management via
CREATE SECRET. - Sanitization: No specific filtering or sanitization procedures for external data content are described.
Audit Metadata