MCP期权策略服务器
此存储库包含用于执行的MCP服务器代码 买入期权和卖出期权。 在…上 克劳德桌面。它提供了一种高效管理期权交易操作的简化方法。
入门指南
先决条件
- 已安装Python 3.8+
uv用于管理依赖关系
设置您的环境
首先,设置您的虚拟环境并安装uv。然后按以下步骤进行:
uv init mcp-server-demo
cd mcp-server-demo然后将MCP添加到您的项目依赖关系中:
uv add "mcp[cli]"要使用uv运行mcp命令:
uv run mcpMCP服务器代码
# server.py
# This script initializes an MCP server for options calculations,
# providing tools to compute breakeven points and payoffs for Call and Put options.
from mcp.server.fastmcp import FastMCP
from options_library import Put, Call
# Create an instance of the FastMCP server
# "Demo_Options" serves as the namespace for the tools
mcp = FastMCP("Demo_Options")
# Define a tool for Call Option analysis
@mcp.tool()
def call_option(strike: float, premium: float) -> dict:
"""
Computes breakeven price and potential payoffs for a Call Option.
Parameters:
strike (float): Strike price of the Call option.
premium (float): Premium paid to purchase the option.
Returns:
dict: A dictionary containing:
- breakeven: Price at which the option breaks even.
- max_profit: Theoretically infinite (since profits scale with the underlying asset price).
- max_loss: Fixed loss, equal to the premium paid.
"""
call = Call(strike, premium)
return {
"breakeven": call.breakeven(),
"max_profit": "infinite", # Unlimited profit potential
"max_loss": call.premium # Limited loss equal to the premium
}
# Define a tool for Put Option analysis
@mcp.tool()
def put_option(strike: float, premium: float) -> dict:
"""
Computes breakeven price and potential payoffs for a Put Option.
Parameters:
strike (float): Strike price of the Put option.
premium (float): Premium paid to purchase the option.
Returns:
dict: A dictionary containing:
- breakeven: Price at which the option breaks even.
- max_profit: Maximum profit if asset price drops to zero.
- max_loss: Fixed loss, equal to the premium paid.
"""
put = Put(strike, premium)
return {
"breakeven": put.breakeven(),
"max_profit": strike - premium, # Profit is capped at the strike price minus the premium
"max_loss": premium # Limited loss equal to the premium
}
运行MCP服务器
mcp install server.pyClaude Desktop配置文件和屏幕截图
{
"mcpServers": {
"server": {
"command": "/Users/arkaroychowdhury/.local/bin/uv",
"args": [
"run",
"--with",
"mcp[cli]",
"mcp",
"run",
"/Users/arkaroychowdhury/Desktop/2025-Portfolio/Code/MCP/mcp-server-demo/server.py"
]
}
}
}