First ever submission, so please be gentle! Here's my train of thought, so feel free to react and tell me where I'm wrong. 
I figured <100 cost wasn't possible because you can't have 2 fixed-length arms reach all 4 slots of glyph of animismus as well as at least 1 side of the same debonder. Between 3 fixed-length arms and 1 fixed-length plus a piston, I figured 3 fixed length would be more efficient because they can work together in parallel. This configuration of 3 arms seemed pretty good, but I didn't test very much. Realizing I could use the extra reagents and flip their orientation also felt like a level up for saving cycles. Finally, I tried to make it a bit more efficient by manually coding the whole sequence, though I think I could have probably done so more efficiently, but ran out of time a bit. 
Prediction: winning submission will be just a bit under 300.