design-taste-frontend-v2-experimental

Pass

Audited by Gen Agent Trust Hub on Oct 22, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches documentation and component resources from well-known technology companies and official public sector repositories (Google, Microsoft, IBM, UK Gov, US Gov).
  • [COMMAND_EXECUTION]: Instructs the use of npx commands for component library initialization (shadcn/ui), representing standard developer workflow for well-known tooling.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from user briefs and external URLs to generate code. Ingestion points: Section 0.A (Briefs, URLs, screenshots). Boundary markers: Section 0.B requires a "Design Read" summary before code generation. Capability inventory: Generation of complex UI components using React and GSAP (Sections 3 and 5). Sanitization: No explicit sanitization or filtering of brief content is defined.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 22, 2026, 11:53 AM