databricks-query

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [NO_CODE]: The skill package is comprised entirely of Markdown documentation and SQL examples. No executable scripts, binaries, or dependency files are present.
  • [PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from external Databricks tables, creating an indirect prompt injection surface. * Ingestion points: Results from SQL queries executed against Databricks tables (SKILL.md). * Boundary markers: The skill does not specify any delimiters or safety instructions to prevent the agent from following commands embedded in the database data. * Capability inventory: The agent is granted the capability to execute complex SQL queries and metadata operations through the DBSQL MCP server. * Sanitization: There is no mechanism described for sanitizing or escaping content retrieved from the database prior to its presentation to the LLM.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 01:09 AM