Why don't you use the OS operator in TPT script along with SED command to replace the character with some meaning ful character. Just need to add one more step and can parse the file.
The C3 87 is UTF-8 encoding for U+00C7, so it would seem that your data is in fact UTF-8 encoded.
Does your TPT script specify USING CHARACTER SET UTF8 for the job?
How are you specifying the TextDelimiter in TPT? Is the script itself in UTF-8? If so, did you use the "-e UTF-8" option on the command line to indicate that fact?
I am facing the same issue,
There is emoji in the data 💯
TPT (tbuild with unicode option) also terminated.
Please let me know if you can make the TPT continue without termination.