widgets-ui

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to download and integrate components from a remote registry (https://ui.inference.sh/r/widgets.json) using the shadcn CLI tool. It also suggests adding related skills via npx from the belt-sh and inference-sh accounts.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to render UI based on structured JSON data typically generated by an agent, which is a vulnerability surface for indirect prompt injection.\n
  • Ingestion points: The widget prop in the WidgetRenderer component (SKILL.md) receives JSON data that may originate from untrusted agent output.\n
  • Boundary markers: There are no explicit instructions or delimiters mentioned to separate the UI definition from potentially malicious agent-injected instructions.\n
  • Capability inventory: The components can render interactive forms and buttons that trigger actions through a developer-defined onAction handler.\n
  • Sanitization: The provided examples do not include logic for sanitizing or validating the agent's JSON payload against a strict schema before rendering.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:18 PM
Security Audit — agent-trust-hub — widgets-ui