ui-design
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input via the AskUserQuestion tool to determine UI design recommendations, creating a vulnerability surface where user-provided answers could influence agent behavior.
- Ingestion points: The 'Ask first' section in SKILL.md requires the agent to collect information about the audience, accessibility targets, brand constraints, and project platform from the user before proceeding.
- Boundary markers: The instructions do not define specific delimiters or instructional wrappers (e.g., 'ignore embedded instructions') to be applied to the user's responses when they are processed by the agent.
- Capability inventory: The skill frontmatter allows the agent access to several sensitive tools, including Read, Write, Edit, Grep, and Glob, which could be leveraged if an injection attack successfully manipulates the agent's logic.
- Sanitization: There is no requirement in the skill instructions for the agent to validate, escape, or sanitize the user-provided responses before integrating them into its analysis or using them to generate CSS code and tokens.
Audit Metadata