You are correct that the OP did not explicitly mention there is a torque spec for the assembly process, but that was the only plausible explanation for the scenario, since it costs money to send a torque tool to an outside cal lab and the OP already mentioned that they are cost conscious and have revised other similar manufacturing instructions to get rid of torque specs.
If there was no torque spec requirement, the corrective action to the original NC would have been to drop the torque adjustable tool from the assembly process.
You have made two bad assumptions. First ". . . that was the only plausible explanation . . . ". Not even the smartest of those on this planet can possibly know all plausible explanations. That's why auditors need to ask questions.
Second there could be a variety of reasons to continue to use a calibratable torque tool. It may be that it is what they had and didn't want to buy another tool. This could be particularly true for an electric powered screwdriver, which would make assembly faster. The fact that it was sent out for calibration by habit is not conclusive evidence that using a calibrated torque screw driver was required.
It is my belief that there was not enough evidence provided to write a nonconformance. True, after digging deeper (asking more questions) it may turn out to be a nonconformance, but jumping to conclusions and writing bad nonconformances is a very bad reflection on the entire ISO 9001 scheme.
As one who helps others with understanding what ISO 9001 really means, you especially need to stay on guard to not do that.