ui-visual-verification
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The file
references/current-guidance.mdcontains two JavaScript utility scripts ('Content-To-Edge Padding Check' and 'Text Overlap Check') intended for execution within a browser environment (via Playwright or DevTools console). These scripts are static templates used to extract layout metrics and do not incorporate untrusted external input into their logic. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze UI data, which serves as a potential attack surface for indirect prompt injection if the analyzed content contains malicious instructions.
- Ingestion points: Rendered UI text, labels, accessibility attributes, and DOM structures extracted by the provided JavaScript scripts in
references/current-guidance.md. - Boundary markers: The instructions do not specify clear delimiters or warnings to ignore instructions embedded within the extracted UI content.
- Capability inventory: The agent is instructed to use various high-capability tools including browser automation (Playwright), mobile testing frameworks (Espresso, XCUITest), and local tool integrations (
device-mcp). - Sanitization: There is no evidence of filtering or sanitization of user-visible text extracted from the UI before it is processed by the agent.
Audit Metadata