better-chatbot-patterns

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • SAFE (SAFE): No malicious code, prompt injections, or unauthorized network operations were identified. The skill focuses on providing reusable implementation patterns for Next.js applications.
  • Unverifiable Dependencies (SAFE): The listed dependencies (zod, zustand, ai, etc.) are standard, well-maintained libraries in the JavaScript/TypeScript ecosystem. No suspicious or unversioned external scripts are executed.
  • Indirect Prompt Injection (SAFE): The skill provides patterns for ingesting untrusted data (FormData) but mitigates the risk by enforcing structured validation using Zod.
  • Ingestion points: Server actions via formData in action-utils.ts.
  • Boundary markers: Implied by the use of structured Zod schemas to separate raw input from application logic.
  • Capability inventory: Database operations (db.update) are mentioned as potential downstream actions.
  • Sanitization: Explicitly uses zod for input validation and superRefine for complex validation rules.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:36 PM