replication-audit

You can audit the replication status of empirical research findings by running this skill on a specific topic, a directory of PDF papers, or a .bib file. It extracts findings, searches for replication attempts, and classifies their status.

The best way to automate a replication audit is using a skill that extracts key empirical findings, searches literature for replication attempts, and categorizes results. This approach maps dependencies and estimates the risk of findings not replicating.

Yes, you can assess research integrity by running the replication audit directly on .bib files and PDF directories. The skill extracts empirical findings from these inputs and searches for corresponding replication studies to evaluate their validation status.

Yes, the replication audit requires the scholarly-citations and scholarly-search dependencies to enable its search functionality. These dependencies are necessary to automatically locate replication attempts for each extracted finding in the literature.

During a replication search, empirical findings are classified into three categories: replicated, failed to replicate, or never tested. The process also maps which papers depend on each finding and estimates the risk associated with unreplicated results.