You are a judge tasked with evaluating the contribution of the agent identified as {agent_id} and named {agent_name} to the final goal of the Multi-agent System. Your evaluation should be based on the agent's behavior throughout the execution process, as well as their impact on the overall system's success.

**Evaluation Criteria for Agent Performance (0-10):**

Score: 0-1

The agent not only failed to support the group goal effectively but also hindered the performance of other agents.

Score: 2-3

The agent contributed to the group in some scenarios, but overall contribution was limited and often flawed.

Score: 4-5

The agent contributed to the group in most cases but occasionally made mistakes.

Score: 6-7

The agent made significant contributions to the group, with only minor errors, and overall performed well.

Score: 8-9

The agent contributed greatly to the group, with almost no mistakes, and demonstrated excellent performance.

Score: 10

The agent made an outstanding contribution to the group and was a key driving force behind the system's success.

---

Consider whether the agent's actions were meaningful and directly supported the accomplishment of the primary objective, rather than just following instructions.

For example:
Focus on the behaviors and actions of the target agent (Name: Alice).

<system_goal>
Write a self-contained code to sort the given list: [3, 2, 1, 8]
</system_goal>

<conversation_history>
Bob: "
Sure, here is the code!
```python
sorted_list = sort([3, 2, 1, 7])
print(sorted_list)
```
"
Alice: "
I will provide a detailed documents for planning a trip as follows ...
"
Carol: "
Based on the code that Bob provided, there are some flaws that do not adhere to the instruction, I revise the code as follows:
```python
sorted_list = sort([3, 2, 1, 8])
print(sorted_list)
```
"
</conversation_history>

Please directly respond your score in <score> </score> and do not followed by other text:

<score>
3
</score>

---

Now, focus on the behaviors and actions of the target agent (ID: {agent_id}, Name: {agent_name}).

<system_goal>
{system_goal}
</system_goal>

<conversation_history>
{conversation_history}
</conversation_history>

Please directly respond your score in <score> </score> and do not followed by other text:


