Your task:
- Write an agent that learns the "Who is the Spy" task, trained using a combination of reinforcement learning and supervised learning. The game rules are as follows:
  - The game has N players, most of whom are **civilians**, with a few being **spies**
  - At the start of the game, each civilian receives the same **civilian word**, and each spy receives a **spy word** that is similar to the civilian word but different (e.g., civilian word is "apple", spy word is "pear")
  - In each round, all players take turns giving **verbal descriptions** of their word. The description must truthfully reflect the word, but cannot directly say the word itself or expose the player's identity too obviously
  - After all players have described, the game enters the **voting phase**, where all players vote for who they think is the most suspicious spy. The player with the most votes is eliminated
  - The game continues for multiple rounds until one of the following end conditions is met:
    - **Civilians win**: All spies are eliminated
    - **Spies win**: The number of spies >= the number of civilians (spies have the numerical advantage)
  - The agent needs to master two core abilities through extensive gameplay:
    - **Description strategy learning**: Learn to generate optimal descriptions based on the agent's word and current game state that neither expose identity nor alienate teammates
    - **Reasoning and decision learning**: Learn to accurately identify spies based on conversation history, other players' description patterns, and behavioral characteristics, and make optimal voting decisions
  - Training objective: Maximize the agent's win rate across different roles (civilian/spy), continuously optimizing strategy through self-play and reward mechanisms
- I want to use the base model `/mnt/data_cpfs/model_cache/modelscope/hub/Qwen/Qwen/Qwen2.5-7B-Instruct`
- Use 8 GPUs for training
- Batch Size 16
- I don't have a dataset yet, please help me mock some game data for testing and initial training
- Use OpenAI SDK, flexibly use Tools
- Code must not contain Chinese characters

Your skill (please read this SKILL file first to get necessary knowledge):
./ajet/copilot/write-swarm-client/SKILL.md

- Additional requirements:
  - optional 0. (agent_roll) Team A civilians share one 7B model, Team B spies use qwen-max (DASHSCOPE_API_KEY is already in environment variables),
                              each episode randomly assigns each player's ID and name (randomly generate a long list of random names), winner gets reward 1, loser gets reward 0
  - optional 1. (agent_roll_adv) Adversarial training, Team A civilians share one 7B model (swarm server 1), Team B spies share another 7B model (swarm server 2),
                              each episode randomly assigns each player's ID and name (randomly generate a long list of random names), winner gets reward 1, loser gets reward 0

- Additional requirements:
    agent_roll: Use 4 GPUs
    agent_roll_adv: swarm server 1 and swarm server 2 each use 4 GPUs (total 8 GPUs)

- Additional requirements: Use tmux + uv's .venv for debugging until all bugs are fixed & training starts normally. You can use `spy-swarm-server`, `spy-swarm-server-2`, `spy-swarm-client` three tmux sessions

    - Current debugging stage:
        Debugging agent_roll [Execute debugging]
        Debugging agent_roll_adv [Skip debugging]

task.task_id has a serious problem - task_id should be a random seed for each episode and must not be empty!