Generate Diverse Test Scenarios Automatically

Using your LLM agent, automatically generate a suite of 10 unique test scenarios for a humanoid robot performing 'object manipulation' tasks in varying environmental conditions (e.g., different lighting, object weights, surface friction, crowdedness). The agent should output a clear, detailed description for each scenario and a corresponding executable simulation script. Demonstrate how the agent explores the test space to ensure diversity and comprehensive coverage of potential challenges.