CodedDataSet for our paper entitled: Agents in the Loop: A Faceted Taxonomy of XR-Supported Remote Collaboration

To provide a systematic overview of prior work, we coded each of the 28 XR-supported remote collaboration papers in our corpus into a structured taxonomy. Table \ref{tab:taxo_table_1} presents the full coded dataset, where each row corresponds to a paper and each column represents one of our selected facets. Specifically, we capture the <b>application domain</b> (e.g., education, industry, social), the <b>XR type</b> (VR, AR, MR, or cross-device), the <b>modalities and graphics engine</b> employed (e.g., Unity, WebXR, haptics, voice, gaze tracking), the <b>extent of AI use</b> (ranging from none to explicit integration of LLMs, VLMs, or other AI models), and the <b>design decisions</b> emphasized by the authors (e.g., representation choices, synchronization methods, or privacy considerations).This tabular representation serves as our coded dataset, making explicit the evidence extracted from each publication. By displaying both non-AI and AI-augmented systems in the same schema, the table highlights recurring design patterns, gaps in evaluation practices, and the gradual integration of AI agents into XR collaboration platforms. The coded dataset also supports reproducibility and enables secondary analysis, such as tracking modality choices across domains or examining the evolution of AI mediation in immersive collaboration.