skills/impertio-studio/mariadb-claude-skill-package/mariadb-core-storage-engines/Gen Agent Trust Hub
mariadb-core-storage-engines
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill identifies and documents the attack surface associated with the
CONNECTandSpiderengines which ingest data from external sources. - Ingestion points: The skill describes how to federate local files (CSV, XML, JSON) and remote databases as SQL tables.
- Boundary markers: The documentation includes explicit instructions to hard-code file paths and avoid user-controlled input in engine parameters.
- Capability inventory: The skill documents the capability to read from the filesystem via SQL, but explicitly restricts this to administrative patterns with proper oversight.
- Sanitization: The skill provides clear remediation guidance, identifying user-controlled file paths as a critical anti-pattern and recommending hard-coding as a safeguard.
- [SAFE]: The documentation is educational in nature and focuses on database architecture and performance. It includes specific defensive guidance to prevent common security misconfigurations related to file access and federation.
Audit Metadata