I do not know of any dynamic block resizing client with consensus from the miners.
When there is some code compiled and running feel free to tell us.
I know. Do you support the concept in principle? As I said, it addresses your scenario under which the 8 MB fork is 8 times more resilient to attack.