AGENT LAB: SKILLS

json-render-react-native

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [Indirect Prompt Injection] (HIGH): The skill is a sink for untrusted UI specifications which can include interactive actions and state mutations.
  • Ingestion points: Renderer component (spec prop) and useUIStream hook (SKILL.md).
  • Boundary markers: Absent; no delimiters are used to wrap or isolate the UI data from the agent context.
  • Capability inventory: State mutation (setState), event handling (Pressable), and dynamic component rendering (registry).
  • Sanitization: Absent; the documentation does not describe any validation or filtering of the incoming JSON logic.
  • [External Downloads] (MEDIUM): The useUIStream export allows the dynamic fetching and execution of UI logic from remote URLs at runtime.
  • [Dynamic Execution] (MEDIUM): The skill interprets a domain-specific logic language for props and visibility (e.g., $state, $cond, $bindState) at runtime.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 12:51 AM