Not only that Yilei annotated with the bug his paper(p37):
"Yilei (April 18) Here is the bug: the amplitude of |φ8.f ⟩ does not satisfy M/2
-periodicity. Another way of
explaining the bug is: the support of |φ8.f ⟩ contains p1...pκ vectors. After domain extension, we should
have got p1p2...pκ · p2...pκ vectors, but as the way |φ8.g⟩ is written, it only contains p1...pκ vectors. So
the expression of |φ8.g⟩ is wrong."
Not only that Yilei annotated with the bug his paper(p37):
"Yilei (April 18) Here is the bug: the amplitude of |φ8.f ⟩ does not satisfy M/2 -periodicity. Another way of explaining the bug is: the support of |φ8.f ⟩ contains p1...pκ vectors. After domain extension, we should have got p1p2...pκ · p2...pκ vectors, but as the way |φ8.g⟩ is written, it only contains p1...pκ vectors. So the expression of |φ8.g⟩ is wrong."
https://eprint.iacr.org/2024/555.pdf