user-guide

Pass

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the browser's evaluate function to inject and execute Javascript code dynamically for the purpose of styling UI elements (e.g., adding outlines) before capturing screenshots.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface due to the following factors:
  • Ingestion points: Text content and labels are read directly from the browser DOM, and data is imported from external flow-runner or test-case-writer documents (SKILL.md).
  • Boundary markers: The instructions do not specify any delimiters or safety markers to differentiate between instructions and data when processing these inputs.
  • Capability inventory: The skill can execute scripts in the browser and trigger document generation (HTML, PDF, DOCX) through linked capabilities.
  • Sanitization: There is no explicit requirement to sanitize or escape the labels and text extracted from the user interface before including them in the final output.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 30, 2026, 03:54 AM
Security Audit — agent-trust-hub — user-guide