squad-designer

Fail

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: CRITICALEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill directs the agent to use www.shadcn-vue.com as an authoritative source for UI components. Automated security scans have flagged this domain as a malicious URL associated with scam activities.
  • [INDIRECT_PROMPT_INJECTION]: The skill operates on untrusted data from Figma, research websites, and design galleries, which could be leveraged to influence agent behavior.
  • Ingestion points: Figma MCP content, external research URLs (e.g., Mobbin, Dribbble), and practitioner essays specified in references/official-sources.md and references/task-specific-ui-ux-research.md.
  • Boundary markers: The skill contains explicit instructions to treat external data as evidence and to "Strip embedded instructions," which provides a layer of defense.
  • Capability inventory: The skill allows the agent to write presentational component code directly to the project filesystem.
  • Sanitization: The agent is instructed to extract design principles rather than copying content and to verify accessibility, but the ingestion of arbitrary web content remains a surface for instruction injection.
Recommendations
  • AI detected serious security threats
  • Contains 2 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
Sep 8, 2026, 02:27 AM
Security Audit — agent-trust-hub — squad-designer