headless-web-viewer

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The script render_url_playwright.mjs fetches geolocation data from the well-known service https://ipinfo.io to configure browser settings like locale and timezone. This is used for realistic web rendering and is documented in the skill instructions.\n- [COMMAND_EXECUTION]: The skill executes a Node.js script that utilizes the Playwright library to launch and control a headless browser. It provides capabilities to save rendered artifacts such as HTML, text, and screenshots to local file system paths specified as arguments.\n- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface due to its core functionality of rendering untrusted web content.\n
  • Ingestion points: Web content is ingested from external URLs via the <URL> argument in the render_url_playwright.mjs script.\n
  • Boundary markers: No explicit delimiters or instructions are used to distinguish between the fetched content and the agent's internal instructions.\n
  • Capability inventory: The skill can perform network operations and write files to the local system via provided paths.\n
  • Sanitization: There is no evidence of sanitization or filtering of the rendered webpage content before it is returned to the agent context.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 03:12 PM