TC.
Regarding the failure in the climb:
If you assume that the FLI parameter is Torque, then by levelling and accelerating you will simply reach a stabilised cruise with less than MCP, if the collective is not touched. The Torque reduction with speed for a fixed collective position is quite dramatic. When the time comes to shut down the engine, the FLM gives no advice other than to "join single engine flight conditions". With the FLI no longer working correctly, this is perhaps not that simple.
My 'policy' is, that if possible, decelerate the a/c to 80kts (Vy) and if required (further) reduce collective to allow the Nr to rise towards the transient limit (390 rpm vs 342 normally). This will give the minimum power demand possible and, allow considerable Nr decay while the 'good' engine spools up to take the load, without drooping below normal (good acceleration with the Arriel). It should also give the best chance of avoiding the use of the OEI HI (30") rating on the 'good' engine, which if used requires the engine to be changed.
Obviously if circumstances dictate otherwise then use it, but it is a rather expensive consequence of a FADEC failure if not absolutely required.
Of course, you can just switch off the engine and deal with the subsequent Nr decay and OEI power useage later, but I (would) prefer to set things up beforehand. Hopefully the mod will be out shortly and this will be a thing of the past, though I don't think the EC-155 has the same history of problems that the 135 has had (well, I know it hasn't).
Whether a run away up or down is possible, is not easy to determine, but I think the likelyhood is remote given the certification requirements. If a failure mode did allow this to happen, I doubt that the engine switch would have any function and in the event of a run up you would have to rely on the overspeed protection shutting it down or, if time permitted, use the fuel cut off lever.