> Sending probes is dangerous. You never know who is going to capture it and figure out where it was sent from.
I'm not sure capture would necessarily be all that likely. Assuming a flyby and not a crash, capturing a probe would require the captor to accelerate to the probe's relative velocity first. If you wanted to bring it back, too, you'd then have to decelerate back into a non-escape velocity with the captive in tow.
Alternatively, you'd need some way of slowing the probe down without touching it, e.g. with light, but doing that would take a rather long time, especially if it wasn't designed as a light sail intended to be slowed down and you don't want to destroy the probe in the process.
Both sound quite nontrivial especially if the probe is travelling at a speed that would be useful for interstellar travel, and if you don't detect the probe a lot in advance, which also sounds nontrivial. If the target world had reached such an advanced technology level that the probe could be detected and deemed interesting early on and captured, all bets would be off anyway.
> Use the energy from the destruction to send a message back on a wide cone that can be intercepted.
Energy from destruction tends not to be very organized. How would you use that energy for sending a coherent directed message that isn't even known in advance (assuming it would actually transmit some of the newly collected data and not just "bye bye")?