wren-generate-mdl
Pass
Audited by Gen Agent Trust Hub on Apr 10, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches a version configuration file from the vendor's official GitHub repository (
raw.githubusercontent.com/Canner/wren-engine/main/skills/versions.json) to verify if the skill is up to date. - [COMMAND_EXECUTION]: The skill relies on the execution of the
wrenCLI and various database-specific drivers (such aspsycopg,google-cloud-bigquery, orsnowflake-connector-python) to perform database discovery and file system operations. - [DATA_EXFILTRATION]: The skill ingests database schema information (tables, columns, types, and constraints). This involves reading metadata from the user's database environment, which is the primary intended function of the skill for project scaffolding.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8) because it processes untrusted metadata from external databases (table names, descriptions) and interpolates them into YAML files. While there are no explicit sanitization instructions, this behavior is standard for database modeling tools.
Audit Metadata