design-taste-frontend
Pass
Audited by Gen Agent Trust Hub on Jul 5, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection through its data ingestion practices.
- Ingestion points: Section 0.A instructs the agent to analyze user-provided signals including URLs, pasted screenshots, and brand briefs to infer design requirements.
- Boundary markers: The instructions do not define delimiters or warnings to ignore instructions that might be embedded within the external data sources.
- Capability inventory: The skill provides instructions for the agent to generate shell commands for package installation (Appendix A) and to use image generation tools (Section 4.8).
- Sanitization: There are no documented procedures for sanitizing or escaping the content retrieved from external sources before it is processed by the agent's logic.
- [EXTERNAL_DOWNLOADS]: The skill references multiple external dependencies and asset providers.
- It recommends installing official packages from trusted vendors such as Microsoft (@fluentui), Google (@material/web), IBM (@carbon), Atlassian (@atlaskit), and GitHub (@primer).
- It references well-known asset providers and CDNs, including Picsum Photos (picsum.photos), Simple Icons (cdn.simpleicons.org), and Shopify (cdn.shopify.com).
- These downloads are neutral and align with the skill's primary purpose of building modern web interfaces using established ecosystems.
Audit Metadata