MCRAG KBS Reproducibility Package

This package accompanies the manuscript "MCRAG: Prototype-Retrieval-Enhanced Mixed-Curvature Learning for Class-Incremental Event Detection", prepared for submission to Knowledge-Based Systems. It contains an archived implementation snapshot, protocol summaries, 60 paper-linked run records, text-free prediction outputs, manuscript table sources, and integrity checksums. The records support inspection of the reported final accumulated-label results, not stage-wise forgetting curves. Exact reconstruction of every historical runtime is not guaranteed. Original corpora, source sentences, trigger text, graph structures, and pretrained model weights are excluded. Third-party resources must be obtained from their providers under the applicable terms. See DATA_ACCESS.md and REPRODUCIBILITY_SCOPE.md. Original software in code/ and verify_results.py is licensed under MIT. Original documentation and derived research artifacts are licensed under CC BY 4.0. The licenses apply to their respective components, not as alternative licenses for the entire package. Third-party rights remain unchanged. See LICENSE_NOTICE.md. Corresponding authors: Kefeng Fan ([e-mail]) and Li Sun ([e-mail]). Funding: National Natural Science Foundation of China, Grant 62672060.