Large language model (LLM) agents turn model-generated decisions into tool calls, API requests, code execution, data access, and transactions. When agent reasoning is manipulated or simply wrong, model-centric guardrails cannot by themselves stop the resulting action. We systematically reviewed runtime-enforcement mechanisms for autonomous and tool-using AI agents in literature published or publicly available from January 2023 to September 2026. Searches of Compendex and the Web of Science Core Collection, followed by deduplication, multi-stage screening, source-grounded eligibility assessment, and targeted snowballing, yielded 58 primary studies, which we coded by enforcement placement, sec... More >

This editorial introduces four papers in Volume 2, Issue 3 of the Journal of Reliable and Secure Computing, covering trustworthy AI governance, secure edge computing, federated fine-tuning, and outsourced LoRA adaptation. More >

Edge nodes are among the most resource-constrained components of computing infrastructure, yet security verification competes directly with application execution for limited compute resources. In Named Data Networking (NDN), each retrieved object carries a producer signature whose verification consumes edge CPU cycles, while existing cryptographic and orchestration studies rarely quantify its impact on service reliability. This paper develops SLED, a secure edge--cloud framework over NDN that incorporates a Security Verification Engine (SVE) into the edge compute budget, explicitly coupling verification occupancy with named-function execution. SLED combines DACSIR, which prices verification... More >

Federated fine-tuning adapts language models to distributed data and is widely adopted as a privacy-preserving alternative to centralized training, yet constrained clients must still store model weights and training states, execute updates, and communicate with a server. This review examines four composable routes---parameter-efficient and quantized adaptation, backpropagation-free adaptation, proxy or submodel adaptation, and split federated adaptation---through a common framework that traces the objects each endpoint retains, exchanges, and discloses. When a client retains the complete base, reducing adapter state leaves a base-storage floor; boundary communication depends on input dimensi... More >

Outsourced model fine-tuning requires reliable task allocation, result validation, and incentives for quality. We present TrustCompute, a compute-sharing platform integrating two-stage adapter verification, reputation-aware scheduling, quality-based payments, and auditable ledger records for LoRA fine-tuning. Our evaluation combines synthetic workloads, real GPU-based fine-tuning, and a local Ethereum Virtual Machine (EVM) environment. Across 440 validation cases, all 400 invalid submissions were rejected and all 40 valid submissions were accepted. With 50% unreliable workers, reputation-aware scheduling achieved 100% task completion and reduced the mean failed-attempt time by 89% relative t... More >

Artificial intelligence is becoming a core engine of corporate digital transformation, but its value depends first on secure, reliable, and accountable data and model infrastructures. As firms combine cloud platforms, edge devices, IoT sensors, digital twins, platform data, and algorithmic decision systems, they also expand the attack surface, privacy exposure, model security risk, and compliance burden. This paper develops a security-aware data and AI governance framework for AI-driven corporate digital transformation. It positions the framework as a unified governance model rather than a narrow extension of data management: data governance controls data classification, provenance, access,... More >

This editorial introduces the four articles published in Volume 2, Issue 2 of the Journal of Reliable and Secure Computing, covering sustainable blockchain architectures, federated learning with large language models at the edge, and cryptanalysis of authentication protocols for maritime and drone environments. More >

Recently, Zahednejad and Gao (2025 Journal of Information Security and Applications, Elsevier, https://doi.org/10.1016/j.jisa.2025.104083,) identified key compromise impersonation (KCI) in both, a two-factor, and a three-factor authentication protocols for IoT devices. They further proposed a lightweight authentication scheme designed to mitigate the identified KCI threats. Subsequently, Zhang et al. (2025, IEEE Transactions on Information Forensics and Security, https://doi.org/10.1109/TIFS.2025.3599678) introduced a chaotic map-based authentication protocol for the Internet of Drones, claiming that their design is resilient to a wide range of security attacks. Earlier, Kumari et al. (2019)... More >