cloud-sql-basics
Pass
Audited by Gen Agent Trust Hub on Jul 16, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs users and agents on how to use the
gcloud sqlCLI to manage infrastructure. This includes creating instances, databases, and setting user passwords. These are standard operations for Google Cloud administration. - [EXTERNAL_DOWNLOADS]: The skill references several official Google Cloud libraries and connectors for Python, Node.js, Go, and Java. All installation commands point to official package registries (PyPI, NPM) or trusted GitHub organizations. These downloads are considered safe as they originate from well-known service providers.
- [PROMPT_INJECTION]: The skill describes an indirect prompt injection surface via the
execute_sqltool in the MCP usage reference. This tool enables agents to fetch and process data from database tables which might contain untrusted content. - Ingestion points: Data returned from SQL queries via the
execute_sqltool (references/mcp-usage.md). - Boundary markers: None explicitly defined in the provided markdown instructions.
- Capability inventory: Full Cloud SQL instance management, database creation, and SQL execution capabilities.
- Sanitization: Relies on the user's implementation of the underlying database connectors and the MCP server's execution environment.
Audit Metadata