api-database-vercel-postgres
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data fetched from external PostgreSQL databases, which serves as an ingestion point for untrusted content that could contain malicious instructions.
- Ingestion points: Database query results are captured in variables like
rowsthrough thesqltagged template orsql.connect()client (SKILL.md, examples/core.md). - Boundary markers: The skill does not define explicit delimiters or instructions to ignore embedded prompts within the retrieved database content.
- Capability inventory: The skill facilitates database read/write operations and connection management. It does not include subprocess execution or arbitrary network requests.
- Sanitization: While the skill correctly mandates the use of tagged templates to prevent SQL injection when sending data to the database, it lacks specific guidance for sanitizing or escaping content retrieved from the database before further processing.
Audit Metadata