sql-queries
Pass
Audited by Gen Agent Trust Hub on Mar 4, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its ingestion of untrusted external data.
- Ingestion points: Processes database schemas and documentation from uploaded files (e.g., database_schema.sql).
- Boundary markers: No explicit delimiters or instructions are provided to ignore commands embedded within the input files.
- Capability inventory: Generates complex SQL queries and executable scripts across multiple database platforms.
- Sanitization: Input data from uploaded files is not sanitized or validated before being used to generate code.
Audit Metadata