ASC Reference Architecture and Framework-Assessment Replication Package

This replication package accompanies the manuscript From LLM Agent Frameworks to Agentic Services: A Service-Oriented Assessment and Reference Architecture. It contains the materials supporting the framework assessment and ATAM-inspired analytical evaluation of Agentic Services Computing (ASC), including: Evaluation protocol and LLM-judge prompt Raw outputs from Gemini 3.6 Flash, OpenAI GPT-5.6 Sol, and Claude Fable 5 Author-validation form and normalized validation record Final ratings and framework-assessment workbooks Non-executable illustrative ASC contract example README documenting provenance, limitations, and reported results The package corresponds to the frozen evaluation snapshot assessed on 27 August 2026. One completed assessment per LLM judge is retained; no repeated-run stability or generalizability claim is made. The contract example is conceptual and non-executable; no parser, validator, or runtime implementation is included.