design
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references font assets from well-known services (Google Fonts) and provides templates that may reference project-specific images. These references are standard for design workflows.
- [INDIRECT_PROMPT_INJECTION]: The skill includes auditing sub-skills (e.g., audit.md) that instruct the agent to analyze external code and content from audited surfaces. This creates an ingestion point for untrusted data. Mandatory Evidence Chain: (1) Ingestion points: Analyzing target page code in sub-skills/audit.md and sub-skills/references/design-critique.md. (2) Boundary markers: Absent. (3) Capability inventory: File reading and UI analysis. (4) Sanitization: Absent. This is a low-severity finding inherent to auditing tasks.
- [SAFE]: The skill consists of markdown instructions and design boilerplate. All embedded code snippets (CSS, HTML, and basic JavaScript for UI interactions) are benign and directly relate to the skill's primary purpose.
Audit Metadata