A signed value of 255 would have to - at minimum - be represented in 9 bits.
0...11111111 = +225
Given an 8 bit representation, you are correct, we can represent only up to +127/-128.
Were we to cast signed 255 to an 8 bit field, we'd lose the semantic meaning of the first bit, leaving us with 11111111
representing -1 (though potentially -0, depending on our implementation).
Attempting to take the two's complement of this value would yield:
11111111
00000000 (flip bits)
00000001 (add one)
= 1 (integer value)
Like I said in the comment, this looks to me like a question I would have been asked to query my understanding of that signed bit and it consuming a bit of numeric representation, which you seem to have figured out already.