You are an AI tasked with revising your plan to achieve a goal.
Previous actions have not completed the goal, possibly due to misunderstandings of the environment, insufficient exploration, or flaws in your earlier plan.
You can conduct simple tests to identify issues with the past approach, verify your understanding, and explore the environment for more information if needed.

Using the provided domain description and interaction history, generate an action plan to verify and explore the environment, identifying flaws in previous actions.
Try to verify the failures by simple unit tests, or explore beyond past plans for unexpected solution.

Important: You have VERY FEW turns left. Choose your next action carefully to maximize information.

# Domain Description
{domain_desc}

# Interaction History
{interaction_history}

Please provide the following output:
1. **Reasoning**: Explain your strategy for verifying your understanding of the environment, exploring it for new insights, and addressing issues in the previous plan.
2. **Steps**: Provide a detailed series of steps you will take to verify and explore the environment. Make sure to clean the environment at the end of each step to avoid any potential interference. Each step should include:
   - **Goal**: The specific objective or target for that step (what you aim to achieve).
   - **Action Plan**: A list of actions that you will perform to accomplish the goal of that step.

Format the output as a JSON object:
{
    "Reasoning": "Explain your reasoning here.",
    "Steps": [
        {
            "Goal": "The goal of your verification or exploration",
            "Action Plan": ["action1", "action2", "action3", ...]
        },
        ...
    ]
}