postgres-nio

Pass

Audited by Gen Agent Trust Hub on Aug 7, 2026

Risk Level: SAFECREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill references the official 'vapor/postgres-nio' library from GitHub, which is a trusted and widely used resource in the Swift development community.
  • [SAFE]: The instructions promote secure communication by detailing how to require TLS for database connections in production environments.
  • [CREDENTIALS_UNSAFE]: Example code snippets use a placeholder password ('secret'). While appropriate for demonstration, users must manage real credentials using secure methods like environment variables or secret management services.
  • [INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface for indirect prompt injection by processing external data from database queries.
  • Ingestion points: Data enters the agent's context through database query results and row decoding operations described in SKILL.md and postgres-patterns.md.
  • Boundary markers: The implementation uses 'PostgresQuery' string interpolation, which effectively separates executable SQL commands from user-provided data using parameter bindings.
  • Capability inventory: The skill facilitates network connections to database hosts and access to local Unix domain sockets.
  • Sanitization: Employs Swift's 'ExpressibleByStringInterpolation' and prepared statements to ensure all external data is treated as non-executable bindings.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 7, 2026, 10:08 PM
Security Audit — agent-trust-hub — postgres-nio