Well, as I wrote, these modifications are probably not a problem. The chance of actually harm anyone or anything are microscopical, but still plausible.
In regards to transmitters, the only way to confirm enhanced antenna functionality without the appropriate test equipment, are the thingies I suggested above.
They will tell most of the truth in regards to the transmitters radiated power and if the antenna modification made it better or worse.
If the antenna works better for the transmitter it should also should have the same effect on the receiver since the equipment work in simplex mode (sends/transmits on the same radio frequency)
But I have a small concern regarding the CPU clock (CPUs software is not my strong suit, however, CPU related radio interference are things that I have worked with before)
CPU hardware has the ability to sometimes interfere with receivers if the CPU clock frequency, or multiples of that comes close to the receive frequency.
But also things such as CPU load (!!!), memory usage and load on other data buses may also make the circuit board to act as an antenna transmitting interfering signals (this is why we have the EMC regulation)
That being said, all of you that have poor coverage, Check your surroundings, do you have any suspicious electronics (i.e. access points etc.)?
Try to move them or temporary switch them off and check if your coverage is better (or worse).
Homey has an internal 1 GHz clock, and that is in the risk zone, frequency is pretty close to 868 MHz.
At the moment I donāt know if the CPU and peripheral circuits have adequate shielding to suppress not interfere (make the receiver hear less good) with ZigBee/868MHz.
I think that I have an old analogue cable TV instrument that can measure this potential EMC field in an controlled environment.
If this is the case, moving Homeys antenna outside and above the casing can actually be a good idea since better balance between transmitter and receiver will be achieved.