Ok the file was xored with 0xaa. So everything is ok now but it uses at least two compression algorithms that I can't recognize at the moment (I didn't check further). In the meantime this is the work-in-progress script, once someone will figure out the two algorithms it will be easy to finish it:
getdstring ZERO 16 get FILES long filexor 0xaa for i = 0 < FILES get SOME_CRC long get OFFSET long get SIZE long get ZSIZE threebyte get ZIP byte if ZIP == 0 log "" OFFSET SIZE elif ZIP == 2 comtype ??? math OFFSET + 4 math ZSIZE - 4 clog "" OFFSET ZSIZE SIZE elif ZIP == 3 comtype ??? clog "" OFFSET ZSIZE SIZE else print "Error: unknown ZIP %ZIP%" cleanexit endif next i