Initial Examinations of CLZ and SB Files

I’ve began taking a closer look at the .clz files that comprise a lot of both AWL and AnWL’s filesystem.

Upon examination with a hex editor, my best guess is that they’re some sort of archive.

The first few hex characters, denoting the CLZ file type header seem to be consistent 0x434C5A.

I determined this because, after looking at commonall.arc.clz, I could find code for a U8 archive (noted by the U8 header 0x55AA382D).  This is likely the  commonall.arc within the  commonall.arc.clz file. However, I’ve found 11 occurrences of this U8 header tag within the CLZ file, so there very well may be 11 different U8 archives (as well as other files with different file headers) within the CLZ archive.

My plan at this point is to basically isolate the U8 archive code, and then save it to a separate ARC file for further analysis.


I’ve also noticed a bunch of SB files in the root/test/Scripts directory of the games.

From what I’ve found online, these seem to be compiled SPC binaries (similar to those in Super Mario Sunshine).

I’m not 100% sure of what these scripts/binaries may do at this time, or if they’re even used at all (as they’re in the test folder).

It seems that an SB Decompiler exists, but now I need to figure out how to compile the decompiler. It seems to require a compiled version of the arookas library to function, but I’m not yet sure how to compile that on a Windows machine.

One step forwards, two steps back.


At this point I’m almost considering emailing Natsume to see if I could get in touch with any of the staff that worked on the original games if they’d be willing to provide any insights into development using these file formats.

But that could also result in them becoming aware of the mod and sending a cease and desist if they deem appropriate. So there is an element of risk in getting them involved.

What are your guys thoughts?