"Meet the Devs" code puzzle

A code puzzle was found in an interview of Chris Horn posted in Black Mesa's official newsfeed on Steam. The interview was part of a series called "Meet the Devs". In the announcement of the interview series, it was stated that an ARG challenge would be embedded in the series.

Solving the outer code layer
An analysis of the code showed that it contained 25 distinct characters, and had a length of 127 characters.

After analyzing the alphabet used in the message, it was discovered that it was a Gold-Bug cipher, which is a simple substitution cipher which uses the following alphabet:

ABCDEFGHIJKLMNOPQRSTUVWXYZ 52-†81346,709*‡.$;?¶]¢:[

A clue hinting at the Gold-Bug cipher can be found in the last sentence above the coded message in the article text: "It’s just not that linear and probably full of bugs..."

The deciphered message is yet another cipher: YGNAEZCRFPDDGLIAGBFRNALWCMSFRWETPGZACAXDRQHFAGDEGILVWTXGBVRBUMZIVLEZDIGEBAAZYMLHYYDLZEBKMCDIEQWFQOFLRCMNSILQGAGOQECNMIPEKVWXGAS

Analysis
Frequency analysis:


 * The ciphertext contains only letters of the alphabet (25 distinct letters), which suggests a classical cipher.
 * There are no occurrences of the letter 'J' in the ciphertext.
 * A Friedman Test was run on the ciphertext with the following results:
 * IC (Index of Coincidence) value: 0.04274 (close to a Vigenère type cipher)
 * Suggested key length: 6.31487 (possibly 7 or 14 characters)
 * Indication of a polyalphabetic cipher
 * A Kasiski test, which looks for repeating ngrams, produced no useful result. This can be easily confirmed by looking at the bigram frequency table above, where we find no bigrams with notably higher frequencies than the rest. This also means that the suggested key length indicated by the Friedman test is likely useless.
 * At 127 characters, the length of the ciphertext not even, which rules out ciphers like Playfair and Four-Square.
 * As indicated above, the IC of the ciphertext is 0.04274. This is in the range of what you would expect for a polyalphabetic or a polygraphic cipher. The expected IC for an English text is somewhere in the range 0.0664–0.0667. A completely random sequence of letters will have an IC close to 0.0385 (1/26) (except when the sequence is too short to produce a flat frequency distribution). A monoalphabetic (substitution) cipher will have an IC the same as that of the plaintext (presumably close to the expected value of an English text), since the cipher doesn't change the frequency distribution. This is also the case with transposition ciphers, since they only permute the order of the letters in the plaintext.

Theories

 * Clues for solving the cipher ("or even a solution") may be hidden in the post with the interview.
 * While "full of bugs" in "It’s just not that linear and probably full of bugs..." seemed to hint at the Gold-Bug cipher, "not that linear" may be a clue to the cipher used for the inner code. The topic of linearity has also come up a number of times in the Discord chat. In a photo posted by Stormseeker of a notebook showing notes for a Hill cipher, which he did "for another project", we can see "non-linear transpositions" being mentioned.
 * The missing letter 'J' in the ciphertext could be an indication that the cipher involves the use of the Polybius square. The references to Arcadia, in the game and in Stormseekers Discord status, may also be clues that support this idea. Polybius, the creator of the Polybius square, was born in Arcadia in ancient Greece.