skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/sql-query-builder/Gen Agent Trust Hub
sql-query-builder
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect injection attacks, specifically SQL injection, due to unsafe data handling practices when generating queries.
- Ingestion points: Accepts project data from external files such as CSV, Excel, and JSON, or direct user input as specified in
instructions.md. - Boundary markers: The instructions and code templates lack boundary markers or warnings to distinguish between data values and SQL commands.
- Capability inventory: The skill requests
filesystempermissions inclaw.jsonand contains logic inSKILL.mdto generate and potentially save database queries. - Sanitization: The implementation in
SKILL.mduses f-strings for query construction (e.g.,f"'{value}'"in_format_value) without escaping single quotes or other SQL control characters, providing a direct surface for SQL injection.
Audit Metadata