Skills
Navigate
skills/quuu/openai-skills/develop-web-game/Gen Agent Trust Hub

develop-web-game

Pass

Audited by Gen Agent Trust Hub on Oct 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes a local Node.js script (web_game_playwright_client.js) to launch a Playwright-controlled browser instance (Chromium) to perform automated interaction sequences on a target URL.
  • [EXTERNAL_DOWNLOADS]: The instructions recommend the installation of @playwright/mcp and playwright, which are well-known and trusted tools for browser automation and model context integration.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill reads state from progress.md and processes content from the web pages it navigates to.
  • Boundary markers: No explicit delimiters or instructions are provided to the agent to treat external game state as untrusted data.
  • Capability inventory: The skill performs file system writes (saving screenshots and JSON logs to the output/web-game directory) and executes JavaScript in the browser context via Playwright's page.evaluate and page.addInitScript functions.
  • Sanitization: The skill does not implement specific sanitization or filtering for the data returned from window.render_game_to_text() or the content of progress.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 23, 2026, 01:40 PM