Note that if I call this "readme" or something like that, the
website does a funny job of displaying it.

The following shows how the masking can be invoked.

You run jwst1pass with some regular finding arguments, such as below.
See the hst1pass ISR documentation for more details on most of them.
The parameters that are not included with hst1pass, or that are particularly
relevant here are:

SHOW_MSK:  this is a boolean that controls whether a mask file is output
           the mask file will just have abcd_cal.fits --> abcd_msk.fits
           2048x2048 simple fits file with no extensions
PSX=       this is a special STDPSX file that I generated by taking my
           empirical PSFs in STDPSF format and concatenating onto them
           the STPSF from Marshall Perrin.  This goes out to 2501 pixels
           radius, thus allowing me to follow saturated star spikes
           This is a 5001x5001 image x1 sampled
NANPATCH+  Default is off
           Since NIRCam has lots of evil NaNs, and NaNs can prevent me 
           from finding stars nearby them, this option will idenfity single
           pixel NaNs and will interpolate over them so that the finding
           isn't affected too much.  I know it's not good, but it's easier
           than teaching my code to measure the pixels around the NaNs, but not
           the NaNs.
FNMSK:     This is the minimum flux of a star that will be used to generate the
           mask
SRCMSK:    This scales up (or down) the mask to do the star rejection; the
           code first finds the saturated stars, measures all stars (sat and unsat),
           then it generates the mask about the stars with Z>FNMSK, then it goes
           through the sources and sees which stars have less flux than mask*SRCMSK,
           these stars are set to 0 flux and -1 for a "n" value in the output list
MKCLX+:    Default is off
           If this is set to true, then the routine will generate a _clx file
           from the _cal file.  It will be identical to the multi-extension _cal
           file, except that:
           1) the error array [2] will add the mask value in quadrature 
              to the error array
           2) the DQ array [3] will contain a flag (256) for pixels that have more
              mask than error in [2]

Here is an example of how it's run -- 

./jwst1pass.2025.09.30_v1f.e \
           HMIN=3 FMIN=100 \
           PSF=STDPSF_NRCA1_F182M.fits \
           OUT=xympqsnN QMAX=0.50 \
           SHOW_MSK+ PSX=STDPSX_NRCA1_F182M.fits \
           REG=xy NANPATCH+ FMNMSK=1000.0 SRCMSK=1.5 MKCLX+ \
           jw01306001001_02101_00001_nrca1_cal.fits