Replication package provides tools for automated software architecture evaluation, enabling thorough assessments.
This replication package accompanies the paper“CAPRA: Scaling Feedback on Software Architecture Deliverables with a Multi-Agent LLM System”. Specifically, the package includes: • A fully self-contained Java-based re-implementation of the CAPRA pipeline built on the Spring AI framework, together with the dependent Python microservices used for document parsing and multi-modal extraction• The CSV export of the Software Engineering (SWE) features extracted via the SALLMA-based workflow and clustered using HDBSCAN for automated Knowledge Base construction• The complete set of 10 CAPRA-generated PDF feedback reports (Teams 06, 12–20)• The full annotation dataset containing binary ratings assigned by two independent evaluators across 8 assessment criteria• One original student architectural deliverable (shared with explicit consent) The package enables full replication of the experimental evaluation described in the paper, including inter-rater agreement computation and Knowledge Base generation. All thresholds, architectural decisions, and pipeline stages described in the paper are preserved.
No takes yet. Share an insight, caveat, or question.
Minin et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: