ui-creator
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on executing local TypeScript scripts (such as
evaluate-ui.tsandinit-component.ts) using thenpx tsxcommand to automate UI assessments and component scaffolding. - [EXTERNAL_DOWNLOADS]: The skill requires the installation of standard web testing and accessibility libraries, specifically
@playwright/testand@axe-core/playwright, to function correctly. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. Evaluation workflows ingest content from external URLs; if a target website contains malicious instructions, they could attempt to subvert the agent's analysis.
- Ingestion points: External web content retrieved via Playwright navigation in
scripts/evaluate-ui.tsandscripts/compare-variations.ts. - Boundary markers: Not present in the scripts processing the external content.
- Capability inventory: Local file system access for writing reports and components, and network navigation capabilities.
- Sanitization: There is no specific sanitization or filtering applied to the external content before the agent interprets the evaluation results.
Audit Metadata