api-baas-neon

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill documents methods for constructing database queries, which presents an inherent surface for indirect prompt injection if user-supplied data is not handled correctly. The skill specifically addresses and mitigates this risk by advocating for best practices.
  • Ingestion points: Query execution examples found in examples/core.md and examples/branching.md.
  • Boundary markers: The skill instructs users to use SQL tagged templates and numbered placeholders to strictly separate query structure from data inputs.
  • Capability inventory: Ability to execute database queries via the neon() function and perform administrative tasks via the Neon REST API.
  • Sanitization: Explicit warnings are provided against the use of sql.unsafe() for user-supplied values, emphasizing the importance of parameterization.
  • [EXTERNAL_DOWNLOADS]: The skill correctly references official resources from the Neon organization, including the @neondatabase/serverless driver and associated GitHub Actions for CI/CD workflows. These are verified as official, well-known service resources.
  • [CREDENTIALS_UNSAFE]: Deterministic scanning identified connection strings in reference.md, which were manually verified as generic documentation placeholders (e.g., [user]:[password]) for user configuration rather than hardcoded sensitive data.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 03:57 PM
Security Audit — agent-trust-hub — api-baas-neon