Midstance knee flexion affects the person's stability. Someone with weak plantar flexors must compensate proximally at the hip, (High-stepping, abducting, circumducting, excessive quad firing, compensatory genurecurvatum, etc.) or must wear an external device (an AFO) that substitutes the force that the plantar flexors ordinarily provide.
An AFO with a dorsiflexion stop can be used to stabilize the knee in extension using GRF control in stance and would also allow knee flexion in swing phase. The degree of correct plantarflexion moment can be dialed in by building the brace in 5-10deg of plantarflexion and adjusting heel height accordingly.
I suppose you could also restrict knee flexion through the use of a KAFO with a locked knee. But the 3PP system which prevents knee flexion in stance phase would also prevent knee flexion in swing phase producing an inefficient walking pattern. The KAFO is a very safe orthosis for sure but, dang, that is one bulky beast of a brace! And Dude still has his Quads, right? Just a 3/5 mmt on the hammy's. The GRF control device (AFO) is more energy efficient but not as safe when compared to the 3PP control orthosis (KAFO) which creates a less energy efficient but safer gait pattern.
A patient might do well with solid ankle AFOs (or a heel-less, anterior cuff style. Then you could include a pressure reducing FO!!) and bilateral rocker sole modifications. If you go this route, the rocker sole mods are the key to controlling the tibia's advancement over the toes through the correct placement of the rocker axis. It will also reduce the plantar pressures that can get pretty high in the FF found with a co-poly full-length footplate.