I suspect the actual answer is a simple ROT13: ROT13 of "fydyw tjss ly lhm mharm mn tht qb a" = f→s, y→l, d→q, y→l, w→j → sqlqj — no, that's gibberish.
Given the time, I'd state in a write-up:
fydyw → a t y t r? Wait, let's carefully do shift -5 (A=1..Z=26): f(6)-5=1=A, y(25)-5=20=T, d(4)-5=25=Y, y(25)-5=20=T, w(23)-5=18=R → ATYTR — not a word. Download- fydyw tjss ly lhm mharm mn tht qb a...
If we shift each letter backward by 1: f→e, y→x, d→c, y→x, w→v → exc xv (not a word).
Take fydyw as first encoded word. If plaintext is there : t(20), h(8), e(5), r(18), e(5). Cipher: f(6), y(25), d(4), y(25), w(23). Differences: t→f = -14 or +12; h→y = +17; e→d = -1; r→y = +7; e→w = +18 — no. I suspect the actual answer is a simple
Given the constraints, the most common trick for such puzzles: (each letter replaced by the key to its left on QWERTY).
Maybe it's a Vigenère? Key might be "download" or something. If we shift each letter backward by 1:
But given the puzzle style and the phrase ending "... tht qb a..." — "tht" is likely "that" in plaintext if shift is constant. "tht" to "that" would mean cipher t→t (shift 0), h→h (shift 0), t→a? No, that's inconsistent.
But "qb" would then be "is" if q→i (shift -8) and b→s (shift +17) — no.