Unable to correctly decrypt. gdc files in Android apks developed using Godot4.6

[rqelove510]

System information

Windows11

Issue description

I found its key using IDA static decompilation project. After game decryption, the file structure, including *. tscn files and some resources, can be seen normally, but. gdc cannot be decrypted. At first, I suspected that the key had been re encrypted by the developer. I repeatedly verified the logic in IDA and confirmed that there was no additional obfuscation logic in the try_open_mack and open_ond_marse methods. I downloaded another game from the developer, which is also an APK. This game was made using Godot 3.5 and surprisingly used the same key. I decrypted it normally using the key I found before, and even unpacked the. gdc file normally.From the output logs, I suspect that it is due to the extensive use of Chinese UTF-8 in the script files. But I cannot be sure

Steps to reproduce

No response

Recovery log

Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (a1)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (81)
Unicode parsing error: Byte 6b is not a correct continuation byte after d0
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (a5)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (a0)
Unicode parsing error: Byte d5 is not a correct continuation byte after f6 b1
Unicode parsing error: Byte 65 is not a correct continuation byte after d5
Unicode parsing error: Byte 57 is not a correct continuation byte after f3
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (b7)
Unicode parsing error: Byte 33 is not a correct continuation byte after ca
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (af)
Unicode parsing error: Byte 4c is not a correct continuation byte after c8
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (bc)
Unicode parsing error: Byte b is not a correct continuation byte after d0
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (88)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (86)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (99)
Unicode parsing error: Byte 71 is not a correct continuation byte after c8
Unicode parsing error: Byte 74 is not a correct continuation byte after f0
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (fa)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (8c)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (98)
Unicode parsing error: Byte ec is not a correct continuation byte after f2
Unicode parsing error: Byte 34 is not a correct continuation byte after ec
Unicode parsing error: Byte 47 is not a correct continuation byte after c1
Unicode parsing error: Byte 49 is not a correct continuation byte after e1 ab
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (91)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (ba)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (fc)
Unicode parsing error: Byte e7 is not a correct continuation byte after db
Unicode parsing error: Byte dc is not a correct continuation byte after e7 b7
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (8e)
Unicode parsing error: Byte 55 is not a correct continuation byte after f6
Unicode parsing error: Byte fb is not a correct continuation byte after c6
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (fb)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (ae)
Unicode parsing error: Byte 5e is not a correct continuation byte after f4
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (8a)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (80)
Unicode parsing error: Byte 56 is not a correct continuation byte after f4
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (93)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (9d)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (aa)
Unicode parsing error: Byte d1 is not a correct continuation byte after cc
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (bd)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (96)
Unicode parsing error: Byte f2 is not a correct continuation byte after c8
Unicode parsing error, some characters were replaced with � (U+FFFD): Missing f2 a3 UTF-8 continuation byte
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (a3)

Link to game download

ad "games.zip" -->

[nikitalita]

I need the link to the full game in order to debug this.

[rqelove510]

I sent it to you via Gmail email gdsdecomp@noreply.github.com

[rqelove510]

By the way, I use 2.5 beta5 for decryption.
Another game developed in 3.5.2 can decrypt. gdc using the same key, and that game also heavily uses Chinese characters. So the problem may still lie in the impact of Godot 4.6's Android compilation template on it. Using 2.5 beta 5 can decrypt some PC games developed by Godot 4.6 normally, but I don't have any more cases to test.

[nikitalita]

I sent it to you via Gmail email gdsdecomp@noreply.github.com

...that's a 'noreply' email. I.e. that e-mail got sent to oblivion.

Post a link to the store page or where it can be downloaded, on this GitHub issue, or I'm closing this. I do not provide assistance with games with encryption unless they're posted publicly.

[rqelove510]

https://www.taptap.cn/app/730132?os=android

[nikitalita]

They’re using a bespoke encryption technique for the GDscript bytecode files. I’ve added the ability to provide your own script to decrypt files are not encrypted using the PCK’s encryption scheme. I’ll PR this shortly.

[nikitalita]

@nikitalita I have a similar question, but I'm not sure if it's exactly the same. Could you take a look at it too? https://store.steampowered.com/app/3609750/_/

GDRE Tools v2.5.0-beta.5
Opening file: Organized Inside Demo/Organized Inside Demo.pck
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (a0)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (ad)
Unicode parsing error, some characters were replaced with � (U+FFFD): Invalid UTF-8 leading byte (b5)

Totally seperate issue. They've messed with the PCK format, but I don't know exactly how.

[nikitalita]

Who commented on this issue previously? I have decided to add custom PCK readers to deal with problems like this and I was actually able to extract Organized Inside successfully; can you please comment here so that you can test this for me?

[nikitalita]

@rqelove510

I have added in support for decrypting files in games like this one in 2.6.0-beta.3.

Below is a script that I wrote for decrypting the scripts. I cannot provide the key, as this is illegal in my jurisdiction, you will have to reverse engineer the key yourself. Tip: Using Ghidra, look for the get_binary_tokens string, and then follow the cross references to the function, it should be pretty obvious once you see it. You can join the GDRE discord for further assistance: https://discord.gg/qhEa6Tgj

class_name CovenantScriptDecryptor
extends CustomDecryptor


# the `key` parameter here is actually the 16-byte key and the 16-byte iv concatenated together
# Go into `Set encryption key` dialog in the 'RE Tools' menu and set the key to the hex string of the key and iv concatenated together
# and also add this script as the custom decryptor script.
func _parse_and_decrypt(file: FileAccess, key: PackedByteArray, is_non_pck_file: bool) -> Dictionary:
	# this is a hacked-in encryption scheme that doesn't get used when decrypting the files inside a PCK
	if not is_non_pck_file:
		return {
			"error": ERR_UNAVAILABLE,
			"length": 0,
			"data": PackedByteArray(),
		}
	var enc_key: PackedByteArray = key.slice(0, 16)
	var enc_iv: PackedByteArray = key.slice(16, 32)

	var result: Dictionary = {
		"error": OK,
		"length": 0,
		"data": PackedByteArray(),
	}
	# file structure is:
	# - encrypted data
	# - padding_bytes (length - 1)

	var data = file.get_buffer(file.get_length() - file.get_position())

	# decrypt the data; uses AES CBC mode
	var ctx = AESContextGDRE.new()
	ctx.start(AESContextGDRE.MODE_CBC_DECRYPT, enc_key, enc_iv)
	var decrypted = ctx.update(data)
	ctx.finish()

	# Check for the "GDSC" magic number at the start of the file; if it's not there, we decrypted this incorrectly and have an invalid key
	if decrypted.size() < 4 or decrypted[0] != 0x47 or decrypted[1] != 0x44 or decrypted[2] != 0x53 or decrypted[3] != 0x43:
		result["error"] = ERR_UNAUTHORIZED
		printerr("Invalid encryption key")
		return result

	# the padding byte count is stored at the last byte of the decrypted data
	var padding_bytes = decrypted[decrypted.size() - 1]
	var length = decrypted.size() - padding_bytes
	decrypted.resize(length)
	result["length"] = length
	result["data"] = decrypted
	return result

func _is_file_nonpck_encrypted(file: FileAccess) -> bool:
	# Only applies to "gdc" files
	if (file.get_path().get_extension() == "gdc"):
		return true
	return false

[rqelove510]

Very helpful, thank you for your effort. Especially the addition of support for custom PCK encryption, which is a stroke of genius.

[rqelove510]

Sorry, I downloaded the latest 2.6beta3 but still cannot decrypt it correctly. I obtained the global key using IDA: 1D * * * It is 32 bytes, and you mentioned that the correct key is a combination of a 16 byte key and a 16 byte IV. Does it have to use dynamic debugging in CBC mode of AES to find the correct KEY? I am familiar with analyzing and debugging Windows projects, but dynamic debugging on Android is very difficult for me. I really hope to get more tips from you.

[nikitalita]

The key and the IV will probably be rendered as multiple 32-bit values in Ghidra. Change the type of the variable that they’re assigned to to a byte array so that you can see the correct order of the bytes.

Also, function that uses the key and the IV after they’ve been assigned will have the key as a parameter before the IV