You describe 4 events (E1 ... E4).

E1: clock A shows 100s
E2: clock B shows 80s

E3: clock B shows 100s
E4: clock A shows 80s

I call the coordinates of frame A (x, y, z, t) and the coordinates of frame B (x', y', z', t').

1) => Events E1 and E2 happen simultaneous for observer A: ##\Delta t_{21} = t_2 - t_1 = 0s##.
But events E1 and E2 do not happen simultaneous for observer B, if they happen at different ##x## coordinates, according to the Lorentz-transformation: ##\Delta t'_{21} \neq 0s##.

2) => Events E3 and E4 happen simultaneous for observer B: ##\Delta t'_{43} = t'_4 - t'_3 = 0s ##.
But events E3 and E4 do not happen simultaneous for observer A, if they happen at different ##x'## coordinates, according to the Lorentz-transformation: ##\Delta t_{43} \neq 0s##.

So there is no contradiction.

The Lorentz transformation for time depends on ##\Delta x##:
##\Delta t '= \gamma (\Delta t - v \Delta x / c^2)##.
If ##\Delta t=0##, then ##\Delta t '= -\gamma v \Delta x / c^2##.

The inverse Lorentz transformation for time depends on ##\Delta x'##:
##\Delta t = \gamma (\Delta t' + v \Delta x' / c^2)##.
If ##\Delta t'=0##, then ##\Delta t = \gamma v \Delta x' / c^2##.