use-style
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's example files reference external resources from well-known and trusted services.
- Fetches web fonts from Google Fonts (fonts.googleapis.com and fonts.gstatic.com).
- Loads the Chart.js library from the JSDelivr CDN (cdn.jsdelivr.net).
- [INDIRECT_PROMPT_INJECTION]: The skill uses user input ($ARGUMENTS or message) to determine which style guide to load from the local file system.
- Ingestion points: User-provided style name enters the agent context (SKILL.md).
- Boundary markers: The skill uses a comprehensive mapping and aliasing logic that acts as an allow-list, effectively mitigating potential path traversal or unexpected file access.
- Capability inventory: The agent is restricted to reading markdown files within the skill's styles directory or specific project override paths (SKILL.md).
- Sanitization: Input is normalized to a known set of aliases (e.g., 'ios' to 'ios-app'), and missing or unknown names default to a help menu rather than direct file execution.
Audit Metadata