neon-serverless

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references external configuration and documentation hosted on a GitHub repository associated with the vendor.
  • Evidence: URL for neon-serverless.mdc in SKILL.md pointing to the neondatabase-labs organization on GitHub.
  • [COMMAND_EXECUTION]: Instructs the agent to install the official Neon serverless driver package using npm.
  • Evidence: Installation command npm install @neondatabase/serverless in SKILL.md.
  • [INDIRECT_PROMPT_INJECTION]: The updateUser function in templates/http-connection.ts provides a pattern vulnerable to SQL injection by ingesting untrusted object keys and values into a database query without proper sanitization.
  • Ingestion points: The updates parameter in the updateUser function within templates/http-connection.ts.
  • Boundary markers: None provided to distinguish between SQL identifiers and untrusted data.
  • Capability inventory: Executes database queries using the neon client.
  • Sanitization: Absent; the function manually builds the SET clause via string concatenation.
  • [DYNAMIC_EXECUTION]: The skill generates SQL statements at runtime through string assembly in its provided templates.
  • Evidence: Implementation of updateUser in templates/http-connection.ts using Object.entries().map() to build query strings.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 05:02 PM
Security Audit — agent-trust-hub — neon-serverless