This occurs in a python script when a cursor comes in contact with a document that contains an invalid utf8 string.

The document does indeed contain an invalid utf8 string. The python driver essentially reads this and crashes the cursor with a fatal error:

bson.errors.InvalidBSON: 'utf8' codec can't decode byte 0xfd in position 1: invalid start byte

I was able to find the affected document and load it in the mongo shell with no errors.

I propose that the python driver if able should handle this in a similar fashion, construct the bson document as best it can, throw a warning if necessary but most important; continue iterating through the cursor.