Issues · cypherstack/generalized-bulletproofs-fix · GitHub

Labels Milestones New issue

When $c_{k,R}$ is $0^n$, $T_i$ for $i$ in $[n' + 1 + (n' / 2) + 1, 2 * n' + 1]$ is effectively the identity

#7

· kayabaNerve opened

on Apr 22, 2026

$W_O$ must be scaled by $x_{j_O}$ in the computation of $\tilde{W}_O$ for correctness

#6

· kayabaNerve opened

on Apr 22, 2026

$\hat{t}$ is dropped from the transcript as the verifier can compute it, except the verifier can't when transformed by the IP proof

#5

· kayabaNerve opened

on Apr 22, 2026

Prover Step 5b omits an inner-product notation present in the original Bulletproofs paper for z W_v, \gamma

#4

· kayabaNerve opened

on Apr 22, 2026

$\tau_i$ is sampled starting from $i = n' / 2$ but 5b iterates it from $i = 1$

#3

· kayabaNerve opened

on Apr 22, 2026

$n_c = 0$ is documented to reduce to BPs, yet there's also a $m \ge 1$ bound

#2

· kayabaNerve opened

on Apr 5, 2026

Commitments are inconsistently indexed from zero/one

#1

· kayabaNerve opened

on Apr 5, 2026