research-supplement-digestion

What problem does it solve?

This skill helps researchers digest AI and deep learning papers' appendices, supplements, and long-form materials to surface reproducibility-critical details, ablation settings, and hidden experiments when the main text lacks clarity for replication.

Core Features & Use Cases

• Structured extraction of data division, training details, optimizer specs, hyperparameter ranges, and evaluation setups from appendices.
• Evidence tracking by locating extra figures, tables, and methodological notes that support key claims.
• Audit-ready outputs that categorize findings into hard_details, probable_assumptions, and still_missing, suitable for downstream planning.
• Workflow integration by aligning discoveries with upstream patterns in references/upstream-patterns.md to guide verification and follow-up research.

Quick Start

Extract reproducibility-critical details from a paper's appendix or supplement to support replication.