Skip to content

Latest commit

 

History

History
102 lines (65 loc) · 3.8 KB

README.md

File metadata and controls

102 lines (65 loc) · 3.8 KB
Raw

r2-ggpack

Suite of r2 plugins to read and manipulate the ggpack data files used by Thimbleweed Park game (if you don't know this game yet, please go and buy it because it rocks).

Here's the video of the talk i gave at r2con2018 about this: https://www.youtube.com/watch?v=36u050C0gzQ

Disclaimer

This is a totally unofficial, unwanted and unrequested work, based on reverse engineering.

Expect a lot of bugs, if you're brave enough to open your data files in write mode with this plugin be sure to work on a copy and make a full game backup first.

Show / Cinema

The above was made in this way:

asciicast

It turns out the same format used to store the ggpack index of files (GGDictionary) it's used also in all the wimpy files, so it's possible to use the =!pDj command to dump those as convenient JSONs:

asciicast

This plugin is also able to deobfuscate the bnut files, which are Squirrel language source code text files containing most of the game logic, so it's possible to view, edit them and have them transparently obfuscated back so to allow game logic modding:

asciicast

Thanks

This wouldn't have been possible without the hints contained in the awesome https://github.com/mstr-/twp-ggdump (even if it turned out that abstraction was wrong)

Why

  • because it's fun
  • to allow more reversing of the game (in the hope to ease the burden of re-writing its engine in the open sources in the future)
  • to enable full mods

What you can do

  • read the contents of the ggpack files
    • browse all files into the ggpack
    • search for content
    • edit them in-place using r2
  • dump the content to files
  • edit the unpacked files with external editors (GIMP, etc)
  • import them again into the ggpack, so that you can just plug that back in the game and see your modifications

How does it work

  • there's an RIO plugin which does most of the job:

    • on-the-fly deobfuscation / obfuscation
    • index directory management

  • a shallow RBin plugin just puts a flag for each packed file

  • a dumb python test suite exists to ensure a minimum level of internal consistency

Installing

Make sure you have radare2 from git before compiling, then just do:

r2pm -i ggpack

or compile it by hand:

git clone https://github.com/mrmacete/r2-ggpack.git
cd r2-ggpack
make install

Usage

Open your data file by invoking radare2 in this way:

$ r2 ggpack://path/to/your/ThimbleweedPark.ggpack1

If you want to modify it (see the disclaimer first):

cp /Absolute/path/to/original/ThimbleweedPark.ggpack1 ./test-twp.ggpack1
r2 -w ggpack:///Path/to/test-twp.ggpack1

There are also a couple of r2pipe scripts:

  • ggimport.py: pass it the path to a file (or a directory), it'll replace the content with the same name in the ggpack, taking into account size differences. If a directory is provided, all the files directly contained into it are imported. Files whose name doesn't match any already existing file in the archive are just skipped, i.e. you can't add new files into the archive but just replace existing ones (for now).
  • ggexport.py: pass it the output directory name (it must exist) and it dumps all resources there

Here's a quick example:

$ggexport=#!pipe python ./ggexport.py
$ggimport=#!pipe python ./ggimport.py

$ggexport dest_dir
$ggimport /path/to/my/NotebookSheet.png