通过调整LLM提示定义AI代理角色

import anthropic

client = anthropic.Anthropic()
response = client.messages.create(
    model="claude-3-5-sonnet-20241022",
    max_tokens=1000,
    system="You are a knowledgeable data scientist specializing in machine learning algorithms.",
    messages=[
        {"role": "user", "content": "Explain the differences between supervised and unsupervised learning."}
    ]
)
print(response.content[0].text)

from openai import OpenAI

client = OpenAI()
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[
        {"role": "system", "content": "You are a helpful assistant specializing in creative writing techniques."},
        {"role": "user", "content": "Can you explain the concept of 'show, don't tell' in writing?"}
    ]
)
print(response.choices[0].message.content)

import os
from openai import OpenAI

class Agent:
    def __init__(self, name: str):
        self._name = name
        self._persona = ""
        self._api_key = os.getenv('OPENAI_API_KEY', '')
        self._model = "gpt-4"

    @property
    def name(self):
        return self._name

    @property
    def persona(self):
        return self._persona

    @persona.setter
    def persona(self, value: str):
        self._persona = value

    def execute(self, task: str) -> str:
        if not self._api_key:
            return "API key not found. Please set the OPENAI_API_KEY environment variable."
        client = OpenAI(api_key=self._api_key)
        try:
            response = client.chat.completions.create(
                model=self._model,
                messages=[
                    {"role": "system", "content": self.persona},
                    {"role": "user", "content": task}
                ]
            )
            return response.choices[0].message.content
        except Exception as e:
            return f"An error occurred: {str(e)}"

# Example usage
if __name__ == "__main__":
    # Create a financial advisor agent
    financial_advisor = Agent("FinancialAdvisorBot")
    financial_advisor.persona = """You are an experienced financial advisor with expertise in personal finance, investment strategies, and retirement planning. Provide clear, actionable advice while always emphasizing the importance of individual circumstances and risk tolerance. Never recommend specific stocks or make promises about returns. Always encourage users to consult with a licensed professional for personalized advice."""

    # Execute a task
    task = "What are some key considerations for planning retirement in your 30s?"
    response = financial_advisor.execute(task)
    print(f"Agent Name: {financial_advisor.name}")
    print(f"Agent Persona: {financial_advisor.persona}")
    print(f"\nTask: {task}")
    print(f"Agent Response:\n{response}")

    # Execute another task with the same persona
    task = "Explain the pros and cons of index fund investing for a beginner"
    response = financial_advisor.execute(task)
    print(f"\nNew Task: {task}")
    print(f"Agent Response:\n{response}")