db-query-patterns
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection by ingesting and processing untrusted local source code and SQL files.\n
- Ingestion points: The skill reads file contents from the local environment using the Read, Grep, and Glob tools specified in the frontmatter.\n
- Boundary markers: The instructions lack explicit delimiters or specific warnings to ignore embedded instructions within the audited data.\n
- Capability inventory: The skill utilizes the Bash tool for pattern searching and database query verification, providing a path for extracted content to influence system state.\n
- Sanitization: While the skill mentions redacting secrets in reported findings, it does not specify sanitization procedures to prevent instruction injection from audited files.\n- [COMMAND_EXECUTION]: The skill uses the Bash tool to perform both static searches and dynamic database verification.\n
- Verification steps involve executing EXPLAIN commands on queries extracted directly from the source code, which could lead to query injection if the analyzed files are maliciously crafted.\n
- The use of the $DATABASE_URL environment variable for connectivity follows standard security practices for secret management.
Audit Metadata