frontend-design

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill defines a robust process with explicit gates for validation and human sign-off, ensuring the agent remains within authorized design boundaries.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided design requirements and content. While this creates a potential surface for indirect prompt injection, it is mitigated by the mandatory manual review process and the use of the repository's established environment.
  • Ingestion points: User-supplied design prerequisites and content data models as defined in SKILL.md.
  • Boundary markers: Human-in-the-loop validation through the /grilling and /wayfinder interactive sessions.
  • Capability inventory: Local file writing for documentation, execution of existing repository task runners, and usage of browser/screenshot tools.
  • Sanitization: Not explicitly implemented in the instructions; relies on user oversight during the design approval process.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 06:58 AM
Security Audit — agent-trust-hub — frontend-design