Here's my thoughts on the points in bold;
4. Higher-quality suggest there's something inferior or deficient with the original caps. Is this down to perception or have you looked at specs or taken measurements that suggests they need replacing? If I'm replacing parts I want to determine what the benefit will be in advance and whether this is worth doing. I think it's a mistake just to replace parts that are perfectly adequate.
5. The same with the tone caps. There's an orange drop madness that I don't subscribe to - a blind belief that every amp made would somehow be better if it had every film cap replaced with orange drops. If you must do this then consider the entire signal path including the coupling caps either side of the PI. The 1000v caps are rated at that because it happens to be a convenient commercial rating for smaller caps ather than one that's necessary in a tone stack. They're really common in power supply applications. Usually pale blue in colour but can vary.
6. As a minimum I like to have a fuse in the HT supply. This is because a common failure mode is an output tube short and this can damage the OT and sometimes the PT, which are particularly vulnerable in lower-priced amps. I would insert an F500mA fuse immedely before the choke and then relocate the B+ feed on the choke side of the fuse.
10. Are you confusing cap rating with value? Sag comes from a combination of resistance and capacitance, that's why tube rectifiers have more sag - there's an intrinsic resistance that increases the time it takes for the main filter caps to charge. There are ways to do this by including series resistance between the rectifier and 1st supply node. You could additionally reduce the value of the PSU caps to give a more 'vintage' value, but at the expense of more hum.
11. Take a look at the London Power articles on power scaling. In the main the preamp side is kept at the original voltage, but the plate and screen voltages are reduced to the power tubes, In advanced circuits the bias is also adjusted. There are some reasonably good output attenuator circuits that are easier, more reliable and less-invasive.
13. I've just added another unused triode half to an amp as an additional series gain stage. Both are simultaneously controlled with a 500k+500K dual pot (only because I didn't have a 1M+1M to hand). It works really well and gives a nice progression from clean to extreme and makes the guitar volume control far more useful. You may need to experiment with attenuation between the stages so that when at maximum the amp is still useable, or make the 2nd stage cathode bypass cap switchable to knock down the overall gain. No NFB loop required.