Hmmm. The original, in fact, does NOT have the 0x0a at the end of the file; however, in process of deleting first part of file up through and including "1.#####", my linux command line tools (or vim?) added one there (unbeknownst to me at the time). I proceeded to solve first puzzle with the 0x0a there at the end in the head-shortened ciphertext, and I got readable plaintext, but my sha256sum does not match yours. If I remove the 0x0a, I get neither readable plaintext nor a matching sha256sum.
This is odd! We can visually troubleshoot this, and see that the first characters following 1.##### are «Cnrtltos » and the final characters at the end of the (original) file are «!niauago». Taking [SPOILER] alternately the first of the first and the last of the last gives us C, o, n, g, r, a, t, u, l, a, t, i, o, n, s, !. If there is an extraneous character inserted at the end, e.g. a newline, the transformation should be spoiled and illegible: C, newline, n, o, r, g, t, a, l, u, t, a, o, i, s, n, space, !.
Dunno if it's helpful for diagnosis, but my 01.txt that yields the expected sha256 sum ...
... is 0x3E1051 bytes long
... starts with <<Congratulations>>, as you already knew it should
... ends with <<IX%7M3+]vW7+zB]{\>>
(It couldn't be line-ending issues, could it? Do you have any 0x0D bytes?)
And, just in case it helps, my original text file (before stripping everything up to 1.#####, and _without_ the spurious 0x0A byte) is 0x3E404D bytes long and ends with << vhuY!niauago>>.
(btw, love this project, thanks for sharing it)