original-design

Pass

Audited by Gen Agent Trust Hub on Jun 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a robust, phase-based methodology for visual identity design. It utilizes standard agent capabilities like web browsing and file management to support its functional goals without any malicious intent.
  • [PROMPT_INJECTION]: The skill identifies an attack surface for indirect prompt injection as it processes user-provided descriptions and external data from design platforms (Category 8).
  • Ingestion points: User responses provided during Phases 0-7 and content extracted from external design websites (cosmos.so, Pinterest, Behance, etc.) during Phase 1.5 and Phase 6.
  • Boundary markers: The skill employs a strictly facilitated, one-phase-at-a-time conversation model and uses a structured Design Language Brief template to delineate user input from system logic.
  • Capability inventory: The agent is instructed to perform web browsing (via Chrome extension or Computer Use), write project files (design-language.md), and invoke other production skills for specialized design tasks.
  • Sanitization: No explicit automated sanitization is described for user input or web content, relying instead on the human-in-the-loop nature of the design session.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 15, 2026, 01:25 AM
Security Audit — agent-trust-hub — original-design