D++: A Declarative Infrastructure Language for Fault-Tolerant, Dynamic Automation in AI-Augmented Developer and Cyber-Physical Workflows.

Modern automation pipelines across cyber-physical frameworks and digital continuousintegration/continuous deployment (CI/CD) lifecycles are increasingly bottlenecked byconfiguration complexity and telemetry fragility. General-purpose scripting environments(e.g., Python, Bash, C++) introduce extensive boilerplate code to compute temporalmetrics, such as evaluating whether a dynamic threshold is sustained continuously oversliding time windows. Furthermore, with the proliferation of artificial intelligence (AI)models generating functional code, structural punctuation syntax (e.g., rigid nesting, whitespacing, symbol enclosures) introduces high rates of deployment failure due to compilerlevel hallucinations.This paper introduces D++, an open-source, declarative, human-readable domain-specificlanguage (DSL) engineered to decouple operational safety policy from low-level executionmachinery. By elevating physical and digital lifecycle boundaries—such as datasanitization, temporal persistence gates, timeout sentinels, and hysteresis trackingfilters—into native syntax primitives, D++ creates a highly robust, fault-tolerantprogramming interface. We validate D++ across two distinct production environments: ahigh-availability smart agricultural automation deployment mitigating physical sensorfailures, and a live, cloud-integrated GitHub Actions DevOps deployment managing flakyautomated software build suites. In both paradigms, we empirically verify that runtimechanges to system safety boundaries are achieved purely through D++ text configurationmodifications without altering the static base interpreter core.