MCP Service Quality: AMQ-Bench English Dataset

This dataset accompanies “Improving MCP Service Quality for LLM Agents: A Benchmark and Empirical Study.” AMQ-Bench contains 269 benchmark instances from 126 open-source repositories for evaluating MCP service Availability, Usability, and Utility. Each instance includes repository information, service construction requirements, a usage task, reference information, and verification materials. The archive provides the full benchmark corpus, field documentation, file checksums, an optional 30-instance subset, and source snapshots for anonymized repository references. The benchmark evaluation framework and AMQ-Wrap implementation are available at https://anonymous.4open.science/r/MCP-Service-Quality-TSC/