Yes, you can safely solder them. I can't see how heating would be an issue unless you cut off the entire wire lead.
IMO, termal blocks, pin headers and all that are just extra failure points. I run the ematch wires through a small hole in the bulkhead and connect them directly to the terminal blocks on the electronics. The small hole and wire can be easily sealed with a bit of "poster tack" (and also easily removed).
https://www.amazon.com/dp/B000BKQDB4/?tag=skimlinks_replacement-20
I don't see pin headers or terminal blocks as failure points; this past weekend I had a recovery failure where the pin header connection stayed intact, and the wires snapped between the connections. I held that particular connection together with one wrap of aluminum foil tape. On Disappearing Act, which had its deployment circuits go through two (apogee) and three (main) double header pin connections in series, the Raven fired both its charges, despite the 220 G boost it underwent.
To be specific, the electricity for the main charge flowed like this (from positive to negative):
lipo -> battery connection adapter pin -> screw switch -> header pin -> another header pin -> terminal block -> long wire down shock cord -> header pin -> Q2G2 -> header pin -> long wire to terminal block -> header pin -> Raven terminal block -> header pin -> battery adapter pin -> lipo.
Despite the horrifying boost that ripped the foil-tape-wrapped apogee charge off of its Q2G2, all 8 pin connections, 6 terminal block connections, and 16 user-soldered connections in series with the main charge held up, as well as all of the other ones for the main charge and powering the Raven itself that I haven't counted. All Q2G2's fired.
The bigger worry for me, especially since I favor the Raven which has a common high, is field usability. User error is more likely when all field-installed connections are made in the same spot. I have my wires and switches and connections that run outside the av-bay pre-arranged and color-coded in order to prevent misidentification: yellow and blue for apogee channels (sun and sky), brown and green for main channels (near the ground and trees), red for high, and black for ground. If I have space, I write actual labels too. Header pins prevent me from mixing up which lead goes to which charge, since the leads are automatically paired together.
The other reason I go through the effort of prepping charges by soldering them to header pins is that I use Q2G2's instead of ematches. The shorter leads on the standard versions can make it impractical to wire directly to an altimeter in larger rockets. For this purpose, header pins are an easy way for a quick, foolproof field connection, putting the thinking work on the preparation side instead of in the field.