Classical mechanics is a good approximation to special relativity, which is a good approximation to general relativity etc. I have heard that if string theory/M-theory is right, then it is not just an approximation to a more accurate theory, but represents the end of the line as a TOE, why is this so? Can this be proved in string theory/M-theory?
I would say that the questions that String/M-theory try to answer are the last ones that our current knowledge of reality allows us to ask.
One may think they are the last ones because they are already indirect, as no obvious experimental fact contradicts the current theories (General relativity and the Standard model of particle physics). Instead of experimental problems, M/String theory addresses the theoretical inconstencies between those theories. This could be seen as far-fetched (not to me), but for sure it will be difficult to imagine more questions afterwards.
Now, one should be careful with the idea of theory of everything for two reasons.
M/string theory progresses, but has not reached yet the point of predicting new facts to allow testing it. It is more in the state of 'consistent set of ideas' than in the state of a complete theory. It might occur that to produce predictions, it has to lower its ambitions to "theory of almost everything".
Past history tells that it already occurred that the Physics community thought collectively that "all was known (but a couple of details)" and actually the "details" led to complete changes (e.g. the creation of quantum mechanics & relativity).
Time will tell!