semantic-view-patterns

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: No malicious patterns, data exfiltration mechanisms, or unauthorized privilege escalation attempts were detected. The skill contains SQL scripts and documentation intended for educational purposes regarding Snowflake Semantic Views.\n- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface in 'Apply mode' where the agent processes user-provided SQL definitions. This risk is inherent to the skill's utility and is managed by the agent's platform-level SQL execution constraints.\n- [COMMAND_EXECUTION]: The skill instructs the agent to execute SQL commands on a Snowflake account to deploy tutorial schemas. This is the primary intended functionality of the skill and is explicitly scoped to the user's provided database and schema context.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 06:35 PM
Security Audit — agent-trust-hub — semantic-view-patterns