Does this seem sensible?
Well, to me... not entirely.
EDIT: I did want to comment on your excellent hand drawn schematic. Nicely done.
Here's the thing. The setup you show will allow for bi-polar shunt behavior of the reverse biased Zener while the forward biased zener will add a .7v drop to forward voltage... in essence leaving you to shunt excess voltage that must first reach a potential of 7.5V before shunting occurs... which already exceeds maximum pin parameters.
I don't think that is ideal.
Attached drawing is how I would do it... Note: it does duplicate what is already IN the AVR chip at the pins...
