snowflake-docs

Pass

Audited by Gen Agent Trust Hub on May 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses sql_execute to perform administrative account operations including installing Marketplace listings (SYSTEM$REQUEST_LISTING_AND_WAIT) and managing databases. These operations are restricted to the intended purpose of setting up the official Snowflake Documentation Cortex Knowledge Extension.
  • [EXTERNAL_DOWNLOADS]: Automatically installs the Cortex Knowledge Extension from the Snowflake Marketplace (listing GZSTZ67BY9OQ4). This is an official Snowflake resource managed within the platform environment.
  • [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection (Category 8) as it retrieves unstructured text data from documentation chunks and incorporates them into the agent's response context.
  • Ingestion points: Data enters the context via the SNOWFLAKE.CORTEX.SEARCH_PREVIEW SQL function in SKILL.md.
  • Boundary markers: The instructions do not define specific delimiters or "ignore" instructions for the retrieved documentation chunks.
  • Capability inventory: The skill has sql_execute capabilities for account-level management and data retrieval.
  • Sanitization: No explicit sanitization of the retrieved documentation text is performed before it is used to generate an answer.
Audit Metadata
Risk Level
SAFE
Analyzed
May 17, 2026, 08:15 AM
Security Audit — agent-trust-hub — snowflake-docs