I thought that there's no reason to mimic sh for this since if you need sh to run a script rc won't work anyway.
So this is just a little syntactic sugar, that for the joy of sl is not compatible with sh, but imho it can also increases the readability of scripts. Indeed I agree with sl that expliitness is an advantage of the current quotation rules.
The idea is to use a single $ to mark the end of variable declarations, so that what's left can't do assignments, and equality is always quoted.
% a=1 echo b=$a
rc: #d/0: token '=': syntax error
% a=1 $ echo b=$a
b=1
% $ eval prefix=$home/foo && echo $prefix
/usr/glenda/foo
% $ eval prefix=$home/foo; echo ./configure --prefix=$prefix
rc: #d/0: token '=': syntax error
% $ eval prefix=$home/foo; $ echo ./configure --prefix=$prefix
./configure --prefix=/usr/glenda/foo
% inf=/dev/random out=/dev/null $ echo dd if=$inf of=$out
dd if=/dev/random of=/dev/null
The mini-syntax should extend till the end of a single command: ; & && and || should stop it.