Comparing their datasheets:
THS3095:
Output current 350 mA [on 50 Ohm equivalent to 17.5V]
[no mention of current protection]
The THS309x incorporates automatic thermal shutoff protection. This protection circuitry shuts down the amplifier
if the junction temperature exceeds approximately 160°C.
THS3002:
Output current 175mA [on 50 Ohm equivalent to 8.75V]
The THS3001 incorporates output-current-limiting protection. Should the output become shorted to ground, the
output current is automatically limited to the value given in the data sheet. While this protects the output against
excessive current, the device internal power dissipation increases due to the high current and large voltage drop
across the output transistors. Continuous output shorts are not recommended and could damage the device.
Additionally, connection of the amplifier output to one of the supply rails (±VCC) is not recommended. Failure of
the device is possible under this condition and should be avoided. But, the THS3001 does not incorporate
thermal-shutdown protection. Because of this, special attention must be paid to the device's power dissipation or
failure may result.
So, strictly it does not, but with the fact you'd have to drive it above 17Vmax, or thus 34Vpp to reach the allowed maximum current, and the fact it has thermal shutdown, and the fact it has two chips in stead of one, I think we are fine.
Edit: thinking about it a bit more, it is not the OpAmp, but the 50 Ohm resistors that we'd have to worry about:
Driving it at 20Vpp Sine or 7.1Vrms, the two build in 100 Ohm resistors have to dissipate 1/2W each.
Driving it at 20Vpp Square or 10Vdc the two build in 100 Ohm resistors have to dissipate 1W each.
I think regardless of 3002 or 3095, it is the resistors that will fail if shorted for a longer time at maximum output. So no change to the short-out protection of the device as a whole. Manual states: "All channels can work more than 60 seconds when the load is short-circuited."
But I don't think I want to run 1W or even 1/2W for a 60 seconds on those resistors
At lower voltages (5Vsq would be 0.01W; 10Vsq 0.25W) we should be fine