Krakatau VS CyberChef

Since Java bytecode uses MUTF-8, which encodes astral characters as surrogate pairs, I've had to implement this conversion in Python multiple times, in Krakatau and Enjarify.

Not sure how applicable this is for your project since my compiler was written in Python, but I used Krakatau.

Here is one that disassembles the bytecodes and reassembles them https://github.com/Storyyeller/Krakatau .

There's an additional problem that unlike regular assembly or LLVM assembly where there's some fully supported standard format, Jasmin is a third party program and different JVM assemblers and disassemblers disagree on so many things. There are also some newer assemblers and disassemblers like Krakatau you could try instead. Krakatau has different syntax than Jasmin or javap.

Krakatau - Java decompiler and disassembler.

The compiler outputs bytecode in a text format, which can be assembled into .class files using the Krakatau assembler. Figuring out how to translate features like nested functions and nonlocals to JVM was pretty interesting, and having access to Java’s standard library made the whole thing much easier than expected.

Then we take the encrypted text and use CyberChef to decrypt it.

Let's go to CyberChef and insert our pieces of evidence.

A parameter was changed from '2AMBCgIQBg' to 'CgIIAdgDAQ%3D%3D' which is just the correct base64 encoding they should have been using the entire time.

I don't think this was a hostile action by Google, I think someone just added better input validation for security reasons and it accidently broke the bad requests they were sending.

https://gchq.github.io/CyberChef/#recipe=URL_Decode()From_Ba...

❗This is indeed the flag, but the text is encrypted with Base64. Usually, the presence of padding character "=" indicates that's Base64 type of encoding (but that's only one of the hints). To decrypt it, we can use CyberChef. Copy-paste the text and we either:

It uses a combination of magic bytes (like the `file` command), entropy analysis and character frequency detection to determine whether an output is likely to be of interest to the user.

The file type mechanism is written here[0]. There's a list of all signatures we detect here[1].

[0] https://github.com/gchq/CyberChef/blob/master/src/core/lib/F...

I think this can be deciphered using CyberChef...