In distribution subs they have thermal ammeters I think they have a 30 minute warm up time so they don't get hammered by large starting currents or faults. I think I'm right in saying that the meters are switched out if circuit forr most of the time I'm you'd you can't leave CTs open circuited so do they just put a resistor across the CTs in are the meters usually left in circuit.
This all depends on where the CT wires go next - the insulation from the meter tails is not the problem, and I can see the metering folk not wanting a box where you cannot see where the tails go as it would be an ideal meter bypass connection point. The CTs for the PV rigs I have seen are insulated singles, and look adequate for alarm/ phone quality wiring.
One saving grace is that the core sizes are small, and step up ratios conspire to not generate the same horrible voltages as the industrial bus bar ten thousand turn units can before saturation.
For example, this brand of inverter and controls which I think are the kind of thin tails kind you refer to, has options to takes a 'high' ratio CT ratio of 2000: 1 or a 'low ratio' one of 300/5 (or 60:1 as we'd call it normally)
To estimate open circuit voltage at core saturation look at the magnetic cross section of that sort of core (Am ~ a couple of cm2 at most ). In a none uniform core this is the point the magnetic field lines all neck down, so the centre leg in a traditional EI core shape, and is where saturation occurs first.
and assuming a core Bsat of 1Tesla
at 50Hz dB/dt when peak B is +/- 1Tesla is ~ 300 tesla per second. But as above here A is only 2/10000 of a square metre, so the volt per turn at saturation is about 600/10000 or 60mV (if you prefer about 15 turns per volt. TPV)
Note we don't normally saturate cores, and for conventional (voltage )transformers at 50Hz, when run backed down from saturation to get sensible core losses, another 'rule of 16' for those who like them, a transformer steel core minimum magnetic area of 16mm by 16 mm is about right for 16 turns per volt..
This contrasts sharply with far bigger dangers from bus mounted CTs with far larger cores (square inch sort of cross sections to go round bigger conductors, so lower TPV, or higher VPT! but also more seriously with greater turns ratios, to bring primary side kA down to something sensible on the secondary).
If operated unloaded these larger cores can generate many hundreds of volts, and in some cases, as well as having scope for killing an unskilled operator, can actually break down the insulation at the terminals and support some impressive arcing. (while doing so, also dropping that fraction of a volt as one "turn" winding on the bus-bar.)
Mike
PS
Areas shaded pink are the limiting core areas in these designs. Ignore the ancient 'inch' notation. Unless you want to memorize 8 turns per volt per 1 inch square core area for 50Hz transformer steels instead....
Needs a double plus good marking for that 
