setup-visual-verify

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill relies on the Playwright framework for browser automation. If it is not already present in the project, the instructions specify using npx playwright, which involves downloading the framework and browser binaries from official Microsoft and NPM registries.
  • [COMMAND_EXECUTION]: The skill installs a template script (templates/screenshot-states.mjs) to the project's scripts directory and provides instructions to execute it via node to perform automated screenshot capture.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface through the processing of external data.
  • Ingestion points: The screenshot script reads UI state configurations from a user-supplied JSON file (e.g., states.json).
  • Boundary markers: No explicit delimiters or instructions are used to isolate the data from the agent's logic.
  • Capability inventory: The skill can write files to the local system (screenshots) and perform network navigation via the automated browser.
  • Sanitization: The script performs JSON parsing but does not sanitize the URLs or DOM selectors provided in the configuration file before passing them to the browser automation APIs.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 03:52 AM
Security Audit — agent-trust-hub — setup-visual-verify