game-qa

Pass

Audited by Gen Agent Trust Hub on Nov 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute various shell commands for setup and testing, including mkdir, cp, and node --test. It also utilizes a standalone iterate client (scripts/iterate-client.js) for rapid feedback loops. These are standard operations for a development and QA tool.
  • [INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface for indirect prompt injection by ingesting data from the game being tested.
  • Ingestion points: Data enters the agent's context through window.render_game_to_text() (JSON state) and browser console errors captured by the iterate client.
  • Boundary markers: The state data is expected in a structured JSON format, which provides some implicit boundaries, though no explicit 'ignore' instructions are mandated for the processing agent.
  • Capability inventory: The skill has capabilities to write to the file system (mkdir, cp), execute tests via Node.js, and perform network requests to localhost for testing the game server.
  • Sanitization: The skill uses JSON.parse() to handle the game state string, which validates format but does not sanitize the semantic content of the state fields.
  • [DYNAMIC_EXECUTION]: The skill makes extensive use of page.evaluate() and page.addInitScript() to execute JavaScript within the browser context. This is the core mechanism of browser-based testing and is used to inspect game globals (window.__GAME__), simulate input, and measure performance (FPS). This is standard and expected for the stated purpose of game QA.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 8, 2026, 03:50 PM