db-keys
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests and processes untrusted metadata from database schemas to generate its reports.
- Ingestion points: The skill reads database schema information via SQL queries (e.g., information_schema) and parses DDL files using the local script
scripts/parse-schema.mjs. - Boundary markers: The instructions do not define explicit delimiters to separate database-provided content from agent instructions.
- Capability inventory: The skill has access to the Bash, Read, Grep, and Glob tools.
- Sanitization: No specific sanitization, validation, or escaping of the ingested database metadata is described or implemented.
Audit Metadata