Artifact evaluation (AE) has become a cornerstone of reproducible research in computer science, with dozens of security and systems conferences running formal AE processes. Yet the resulting data remains fragmented, as outcomes are scattered across venue-specific, inconsistent websites, lacking machine-readable metadata and cross-venue aggregation. This prevents the community from answering fundamental questions about AE adoption trends, institutional participation, evaluator workload and retention, and AE sustainability at scale. It also breaks the creation–evaluation–reuse loop: no search by, e.g., topic, exists across security and systems artifacts, limiting artifact discovery and reuse.
We present ReproDB, an open-source automated pipeline that scrapes AE results, committees, author metadata, and repository statistics from 13 security and systems conferences (2017-2026) and homogenizes them into a unified dataset. Building on this foundation, the platform enables three capabilities: (1) the first cross-venue analysis of AE, revealing AE health challenges and open-science policies as the strongest lever for participation; (2) a combined metric that captures both artifact creation and AE committee service, visualizing reproducibility labor; and (3) the first cross-venue artifact search engine, closing the creation–evaluation–reuse loop by enabling discovery by topic, author, institution, or venue.
Accessible at: http://reprodb.github.io