成功标准

此示例展示了如何设置代理的成功标准。

代码

cookbook/agent_concepts/other/success_criteria.py

from agno.agent import Agent
from agno.models.google import Gemini
from agno.tools.thinking import ThinkingTools

puzzle_master = Agent(
    model=Gemini(id="gemini-2.0-flash"),
    tools=[ThinkingTools(add_instructions=True)],
    instructions="You are a puzzle master for small logic puzzles.",
    show_tool_calls=False,
    markdown=False,
    stream_intermediate_steps=False,
    success_criteria="The puzzle has been solved correctly with all drinks uniquely assigned.",
)


prompt = """
Create a small logic puzzle:
Three friends—Alice, Bob, and Carol—each choose a different drink from tea, coffee, and milk.
Clues:
1. Alice does not drink tea.
2. The person who drinks coffee is not Carol.
Ask: Who drinks which beverage?
"""

puzzle_master.print_response(prompt, stream=True, show_reasoning=True)

用法

1. 创建虚拟环境

   打开 Terminal 并创建一个 python 虚拟环境。

   Mac

   python3 -m venv .venv
   source .venv/bin/activate
   
2. 设置您的 API 密钥

   export GOOGLE_API_KEY=xxx
   
3. 安装库

   pip install -U google-generativeai agno
   
4. 运行 Agent

   Mac

   python cookbook/agent_concepts/other/success_criteria.py