Anonymous credentials (ACs) are secure digital versions of credentials that allow selective proof of possession of encoded attributes without revealing additional information. Attributes can include basic personal details (e.g., passport, medical records) and also claims about existing attributes (e.g., age > 18), which can be revealed without disclosing any concrete information. However, embedding all possible claims in a credential is impractical. To address this, we propose verifiers defining policies as high-level programs executed by holders on their credentials. We also propose making the proofs designated verifier to prevent the misuse or leakage of sensitive information by dishonest verifiers to any unwanted third party.

To detect runtime attacks against programs running on a remote computing platform, Control-Flow Attestation (CFA) lets a (trusted) verifier determine the legality of the program’s execution path, as recorded and reported by the remote platform (prover). However, besides complicating scalability due to verifier complexity, this assumption regarding the verifier’s trustworthiness renders existing CFA schemes prone to privacy breaches and implementation disclosure attacks under “honest-but-curious” adversaries. Thus, to suppress sensitive details from the verifier, we propose to have the prover outsource the verification of the attested execution path to an intermediate worker of which the verifier only learns the result. However, since a worker might be dishonest about the outcome of the verification, we propose a purely cryptographical solution of transforming the verification of the attested execution path into a verifiable computational task that can be reliably outsourced to a worker without relying on any trusted execution environment. Specifically, we propose to express a program-agnostic execution path verification task inside an arithmetic circuit whose correct execution can be verified by untrusted verifiers in zero knowledge.

In the race toward next-generation systems of systems, the adoption of edge and cloud computing is escalating to deliver the underpinning end-to-end services. To safeguard the increasing attack landscape, remote attestation lets a verifier reason about the state of an untrusted remote prover. However, for most schemes, verifiability is only established under the omniscient and trusted verifier assumption, where a verifier knows the prover’s trusted states, and the prover must reveal evidence about its current state. This assumption severely challenges upscaling, inherently limits eligible verifiers, and naturally prohibits adoption in public-facing security-critical networks. To meet current zero trust paradigms, we propose a general ZEro-Knowledge pRoof of cOnformance (ZEKRO) scheme, which considers mutually distrusting participants and enables a prover to convince an untrusted verifier about its state’s correctness in zero-knowledge, i.e., without revealing anything about its state.

As lightweight embedded devices become increasingly ubiquitous and connected, they present a disturbing target for adversaries circumventing the gates of cryptography. We consider the challenge of exfiltrating and locating cryptographic keys from the run-time environment of software-based services when their software layout and data structures in memory are unknown. We detail an attack that can, without affecting the system’s operation, exfiltrate keys in use promptly by leveraging the strong causality between transceivers and keyed cryptosystems (authentication, authorization, and encryption). We then propose how to effectively and efficiently reduce the key material’s search space from a batch of stackshots (stack extractions) by leveraging the stack’s innate composition, which, to the best of our knowledge, is the first method to systematically infer and reduce the search space of semi-arbitrary keys. We instantiate and evaluate our attack against MSP430 micro-controllers.