I want to get results for only working days to get the Prophet forecast data

The Profet algorithm page contains the

Prophet is especially useful for data sets such as:
·Includes detailed historical observations (every hour, daily, weekly) over a long period of time (monthly, annual)
·Multiple strong seasonality
·Includes previously recognized, important but irregular events
·No data points or large missing values
·There is a nonlinear growth trend approaching the limit

From this point of view, it is considered appropriate that the actual data used for the analysis should be daily sales, including non-working days.
Why don't you add zero sales for non-working days (days not included in the test results data) to the test results data and use date-sorted data to predict?

Prophet seems to use a Fourier analysis method, so continuous input values (performance data to be analyzed) should be desirable.

I think it's honest to get a year's worth of data prediction from one year's worth of data (365 days).
Also, I think it is reasonable to assume that there will be no sales on non-working days.

The following page is in English, but I think it will answer this question.

How to exclude weeks (Saturday & Sundays) from fb Prophet Issue#1291 ·facebook/profet

This will discard the weekend date in df:

df_no_weekdays=df['ds'].dt.dayofweek<5]
# For example:
# m = Profit()
# m.fit(df_no_weekdays)

The following pages may also help you solve the problem.

Eliminate weeks?·Issue#323·facebook/profet

This only creates a forecast date for business days:

horizon=200
future=m.make_future_dataframe(
    periods=horizon, 
    freq = 'B'
)

'B': Every business day (Monday - Friday)

(I'm sorry, but my native language is English.)