A behavioral risk-aware access control framework for secure digital legacy systems
Digital legacy systems must decide, without the account owner present to confirm intent, whether a request for emergency access to sensitive testamentary documents is legitimate. Existing approaches address this problem only partially: platform-specific legacy-contact features are not interoperable and require no independent corroboration, attribute- and role-based access control models regulate access during active use but are not designed for posthumous or inactivity-triggered release, and machine-learning-based anomaly detection improves suspicious-activity detection at the cost of data and computational requirements that are impractical for a lightweight, single-tenant application. This paper proposes and evaluates a behavioral risk-aware access control framework that unifies four mechanisms within a single architecture: (i) a rule-based behavioral risk-scoring engine that continuously derives a 0-100 risk value from failed-login and high-risk-action signals; (ii) a three-stage, checkpoint-based inactivity/liveness-detection algorithm that escalates from confirmation e-mails to an automatic emergency-access trigger; (iii) a threshold (2-of-3) multi-party authorization protocol that releases access only once independent trusted contacts submit single-use cryptographically random codes; and (iv) role-based access control combined with AES-256-CBC document encryption and full audit logging. We instantiate the framework in a working prototype, LegacyVault, built on Node.js, Express.js, and MySQL, and evaluate it through functional, performance, and security testing aligned to an explicit threat model covering credential compromise, trusted-contact collusion, and insider access. Testing shows that all implemented controls (authentication, encryption, threshold approval, inactivity escalation, and audit logging) operate as designed, with core operations completing in 1.2-7.0 seconds on a local test deployment. We discuss how the framework's design compares with role/attribute-based access control, single-party legacy-contact tools, and blockchain and ML-based alternatives, and we report the scale limitations of the present evaluation transparently rather than overstating a small, single-environment test as a production-grade security guarantee.