skills/vercel-labs/json-render/vue/Gen Agent Trust Hub

vue

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [External Dependencies]: The skill requires the installation of external Node.js packages including @json-render/vue, @json-render/core, zod, and vue. These libraries provide the core functionality for the UI rendering and state management features described in the skill.
  • [Indirect Prompt Injection Surface]: The skill is designed to process JSON specifications to render UI components (<Renderer :spec="spec" />). As the instructions note this can be used for 'rendering AI-generated specs', there is an inherent attack surface where the processed data could attempt to influence the agent's behavior or trigger unintended component actions if the source of the JSON is untrusted. This is a standard consideration for data-driven UI renderers and is mitigated by the use of a predefined component registry and catalog which limits the available actions and logic.
  • [Dynamic Logic Evaluation]: The UI system supports dynamic expressions such as $state, $bindState, and $computed. These allow logic to be embedded within the JSON specifications. While this enables flexible UI behavior, it requires the underlying library to safely parse and execute these expressions to prevent arbitrary code execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 04:00 PM
Security Audit — agent-trust-hub — vue