Why is my Google Earth Engine NDVI showing up with reversed values way outside the bounds of -1 to 1?

Question

I am having a lot of issues with creating an NDVI raster using Google Earth Engine. Everything else works- just not the actual NDVI values. When I export the layer to ArcGIS, it shows up with values of -1000 to 8000 or with no variation at all. I am trying to calculate zonal statistics of NDVI in different land areas in Montana.

//Sources: https://igis.ucanr.edu/Tech_Notes/EarthEngine_NDVI/
//https://medium.com/@moraesd90/creating-monthly-ndvi-composites-sentinel-2-on-google-earth-engine-a5c2d49bc9ca
//https://mygeoblog.com/2016/10/26/sentinel-1-2-for-high-resolution-landuse-mapping/ 
//https://developers.google.com/earth-engine/classification


var geometry = geometry22;
//define study area and set date ranges
var study_area = geometry;
var date_start = ee.Date('2023-06-01');
var date_end= ee.Date('2023-06-30');

//center map
//var long = ee.Number(study_area.centroid().coordinates().get(0)).getInfo();
//var lat = ee.Number(study_area.centroid().coordinates().get(1)).getInfo();
//Map.setCenter(long,lat,11);

//define collection
var S2 = ee.ImageCollection("COPERNICUS/S2_HARMONIZED")
                .filterBounds(study_area)        
                .filterDate(date_start.advance(-1,'year'),date_end);

// cloud function to remove clouds
var cloudfunction_ST2 = function(image){
  //use add the cloud likelihood band to the image
  var quality = image.select("QA60").unmask();
  //get pixels above the threshold
  var cloud01 = quality.gt(0);
  //create a mask from high likelihood pixels
  var cloudmask = image.mask().and(cloud01.not());
  //mask those pixels from the image
  return image.updateMask(cloudmask);
};
 
// remove the clouds
var ST2_nocloud = S2.map(cloudfunction_ST2);

//create NDVI funtion

function addNDVI(image) {
  var ndvi = image.normalizedDifference(['B8','B4'])
                  .int16();
  return image.addBands(ndvi.rename('ndvi'));
}

//select bands of interest (NDVI only)
var ST2_nocloud = ST2_nocloud.map(addNDVI);
var ST2_nocloud = ST2_nocloud.select('ndvi');

//clip collection to study area
                    
var S2_clipped = ST2_nocloud.map(function(img) {
  return img.clip(study_area).reproject({
    crs: 'EPSG:4326', // Set your desired CRS
    scale: 1000 // Set your desired scale
  });
});

                    
                    
//create monthly composites
//create list of months from 1 to 12
var months = ee.List.sequence(1, 12);//separate by years
var years = ee.List.sequence(date_start.advance(-1,"year")
                                       .get("year"),
                             date_end.get("year"));
                             
                             
//compute median value per month
var composite = years.map(function (y) {
     return ee.ImageCollection.fromImages(months.map(function (m) {
      return S2_clipped.filter(ee.Filter.calendarRange(y,y,'year'))
                       .filter(ee.Filter.calendarRange(m,m,'month'))
                       .median()
                       .set('month',m, 'year',y);
    }))
  });
  
  
  
//decompose list of img collection into list of images
function decomposeList(l) {
  return ee.ImageCollection(l).toList(12)
}var list_imgs = composite.map(decomposeList).flatten();

//set system id and system index according to year and month
//set system id and system index according to year and month
function renameImages(img1) {
  var img = ee.Image(img1);
  var value = ee.Number(img.get('year')).format('%04d')
          .cat('_').cat(ee.Number(img.get('month')).format('%02d'));
  return ee.Image(img.set('system:index', value, 'system:id',value))
}var list_imgs_renamed = list_imgs.map(renameImages);

//convert from list to img collection
var decomposed_collection = ee.ImageCollection(list_imgs_renamed);



//convert from img collection to multi-band single image
var singleBand = decomposed_collection.toBands();

//creating NDVI color palette
var NDVIpalette = ['FFFFFF', 'CE7E45', 'DF923D', 'F1B555', 'FCD163', '99B718',
                    '74A901', '66A000', '529400', '3E8601', '207401', '056201',
                    '004C00', '023B01', '012E01', '011D01', '011301'];

//add june composite to map
Map.addLayer(singleBand,{bands: '2023_06_ndvi', min: -1, max: 1,
                       palette: NDVIpalette}, 'ndvi june 2023');
            


var exportRegion = geometry22;


// Define the export parameters
var exportParams = {
  image: singleBand,
  description: 'June_2023_NDVI', // Set your export description
  scale: 1000, // Set your desired scale
  region: exportRegion,
  maxPixels: 1e10, // Set the maximum number of pixels
  fileFormat: 'GeoTIFF', // Set the file format to GeoTIFF
  formatOptions: {
    cloudOptimized: true
  }
};

// Export the image to Google Drive
Export.image.toDrive(exportParams);

I changed the part where it is scaling to 10000, but then all pixels are the same color, with no variation in NDVI, because that appears to be the main culprit. Also, even if I divide the values by 10,000 in ArcGIS, the values seem to be reversed, i.e. heavily forested areas are the negative ones, and the values are just not what I would expect.

Any ideas for what else could be going wrong? I am very unfamiliar with Google Earth Engine and it is not in my preferred coding language.