Paper 2025/161 Original: eprint.iacr.org/2025/161 ↗

Stronger Security for Blind Signatures with Attributes

Abstract

Anonymous Attribute-Based Credentials (ABCs) let users authenticate by proving only the attributes required by a policy. In many applications, single-show ABCs are particularly attractive because they can be instantiated from Blind Signatures with Attributes (BSAs) and showing reduces to a signature check, yielding low overhead. However, most existing Blind Signatures do not allow for attribute embedding. We provide a generic transform that lifts a broad class of two/three-move blind signatures to BSAs. We furthermore identify issues with prior definitions of blindness and secure showing of BSAs, particularly related to the selective reveal of attributes, and we upgrade these definitions. Our transform uses Merkle commitments and a parallel cut-and-choose to enforce attribute consistency and achieves concurrent one-more unforgeability, our stronger secure-showing notion, and the upgraded blindness. Finally, we prove that the well-known construction of Anonymous Credentials Light (ACL) satisfies the stronger notions without modification (in ROM+AGM). Taken together, our results consolidate the definitional landscape for single-show ABCs with attributes and provide a modular route to practical constructions.