The torque fluctuations in the FDR data are interesting.
NH2 looks like it went to roughly 80%, if I'm reading it correctly, which must be well above idle, but possibly low enough to cause the torque to fall below the threshold for the auto-feather. With the engine still running, above idle, the torque would increase as the prop feathered. Would the auto feather cancel if the torque rose above its threshold? If so, the torque would fall again as the prop started to unfeather, and you could end up with an oscillating torque.
The #1 torque fluctuations have got me completely baffled though, as the prop shouldn't be cycling after the #1 condition lever was put to cut-off.
I suspect there is a lot more to this story to come. The indications in the cockpit may not have been as clear as some here seem to assume.