If you are using Python here anyway, why are you not implementing the entire operation in Python?
#!/usr/bin/python3
import fileinput
for line in fileinput.input():
print(line.replace("\u200e", ""), end="")
Demo: https://ideone.com/5dV285
If you insist on a one-liner, try with Perl instead of sed
:
perl -CSD -pe 's/\x{200e}//g'
Demo: https://ideone.com/JAQGu0
If you can get the proper UTF-8 encoding of the character into a variable, removing the square brackets should work trivially with most sed
implementations.
char=$(python3 -c 'print("\u200e")')
echo "be" | sed "s/$char//g"
Demo: https://ideone.com/TrvVJj
Tangentially, avoid upper case for your private shell variables.