Finance Tools Code MCP Server

This folder contains a Code MCP server template for finance and basic quantitative analysis workflows in SAS Retrieval Agent Manager (RAM).

The server exposes tools for:

• Fetching current market prices for one or more ticker symbols
• Fetching recent price history for a symbol
• Running simple linear regression predictions
• Computing summary statistics for numeric inputs
• Performing basic arithmetic helpers

Template Files

• run.py: MCP tool implementations and tool definitions
• requirements.txt: optional third-party dependencies for this server template

Tools

fetch_current_price(ticker_symbols: list[str]) -> dict[str, float]

Returns the latest available market price for each ticker symbol in the input list.

• Input: list of ticker symbols (example: ['AAPL', 'MSFT'])
• Output: dictionary mapping ticker symbol to current price
• Behavior note: if a live price is unavailable, the tool currently returns -1.0 for that symbol

fetch_price_history(ticker_symbol: str)

Returns the recent historical price data for a ticker symbol.

• Input: single ticker symbol
• Output: recent price history (30 day period)

predict_linear(x: list, y: list, x_new: list) -> dict

Fits a linear regression model and predicts values for new inputs.

• Input:
  • x: training inputs
  • y: training outputs
  • x_new: values to predict
• Output:
  • predictions
  • coef
  • intercept
• Validation: returns an error object if x and y have different lengths

get_summary_statistics(values: list) -> dict

Computes descriptive statistics for numeric values.

• Input: list of numbers
• Output includes:
  • count, mean, median, std, min, max
  • quartiles (25%, 50%, 75%)
  • skew, kurtosis
• Validation: returns an error object for empty or non-numeric input

add(a: float, b: float) -> float

Returns the sum of two numbers.

multiply(a: float, b: float) -> float

Returns the product of two numbers.

Environment Variables (optional)

No environment variables are required for this example.

Notes

• This template uses third-party libraries (yfinance, numpy, scikit-learn) in run.py.
• Ensure those dependencies are available in the template environment.
• Market data retrieval depends on external data availability and network access.