CoT+ (Chain-of-Thought Extended): Reproducibility Workflow

This document serves as the companion reproducibility package for the manuscript\emph{CoT$^{+}$ (Chain-of-Thought Extended): A Reproducible Multi-Reasoning Prompting Architecture for Deterministic PLC Code Generation}.Its purpose is twofold: (i) to provide a complete, implementation-level specification of the proposed CoT$^{+}$ prompt protocol, and (ii) to enableindependent researchers to replicate all reported comparisons against the baseline prompting strategies (CoT, PoT, CoC, and CCoT) on arbitrary external PLC datasets.

To support transparent replication, the document consolidates the full prompt cards, fixed protocol obligations, and artifact definitions(\texttt{PHASE 0 / [TASK\_SPEC]} and \texttt{S1--S4}) used throughout the study, together with explicit acceptance rules(e.g., symbol-contract freezing, verification verdicts, and traceability requirements). In addition, it provides step-by-step reproducibility workflowsfor each evaluation layer employed in the manuscript: BLEU-based lexical determinism (including normalization and deterministic BLEU calculation),LLM-in-the-loop validation with five independent validator models (LITL), and human-in-the-loop expert assessment (HITL), including the inter-raterreliability (IRR) procedures based on Cohen's $\kappa$ and Krippendorff's $\alpha$ with a fixed stability gate.

Overall, this documentation is designed to function as a standalone replication manual: by keeping the protocol components fixed and substituting onlythe task instances, an external researcher can regenerate the intermediate artifacts, recompute the reported statistics, and verify the prompt-inducedeffects of CoT$^{+}$ relative to the baseline suite under controlled and auditable conditions.