First, sorry for the slow response.
I haven't ran it on AGOT in a few (of their) versions. Sometimes minor changes or edge cases that are valid for CK2 bork the script. For AGOT, this is usually Dothraki stuff.
I downloaded 0.9.5, generated it, and ran the validator. Looks like it was generating some weird code.
Minor stuff (fixed):
I forgot to switch over "build_cost" to "gold_cost" in the faux buildings, but that shouldn't have caused any issues.
I was generating emptyFROM ={} blocks when there was no prestige cost for a building, but that should've been ignored.
ca_asoiaf_move_camp has a weird potential block, I'm considering just excluding it; it probably doesn't make sense to autobuild it anyway.
Something I did borked the validator, I guess. Or maybe there's a bug.
With ca_asoiaf_move_camp excluded, it works, with it included, it has your error. There's a block called exception_list in the code for excluding certain buildings by name. I don't think excluding this particular building will matter, one way or another, but I'll take a look at...translating its potential block.
owner = { character = FROM }
I think this translates semantically, to, the province owner owns the settlement. But my naive script translates it to the event location block as
owner = { character = owner }
Which isn't the same thing. My hack was to scope to the province owner instead, which in this case, is apparently wrong enough that it causes CK2 to crash. I'm not sure why I was doing that. (Edit: On reflection, I wrap the translated potential blocks in a location scope as well, which might be unnecessary now...) Changed it to controller and everything appears to work. Excluding that one building also codes around the issue.
I am going to run a few more tests, but will upload a fixed script soon, as well as a generated version for 0.9.5 to moddb.