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

chat

Pass

Audited by Gen Agent Trust Hub on Nov 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • SDK Integration and Dependencies: The skill provides instructions for utilizing standard Node.js packages such as @json-render/chat and the Vercel AI SDK (ai). These dependencies are consistent with the skill's stated purpose of building AI-driven chat applications and originate from recognized development ecosystems.
  • Data Ingestion Surface: The skill describes a workflow where user-supplied text (message.text) is processed by an LLM to generate UI specifications. This creates a standard surface for indirect prompt injection common in AI applications. The skill addresses this by providing structured validation through the toCard function and CardSpecError reporting, which allows the system to identify and repair malformed specifications.
  • Restricted Rendering Capabilities: The skill defines a set of standard components (e.g., Text, Fields, Table) that map to platform-specific UI elements like Slack Block Kit. By using a predefined schema and limits on elements (e.g., maxElements), the skill maintains boundaries on how external data is represented, reducing the risk of UI-based exploits or platform degradation.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 5, 2026, 10:46 AM