Ai Research Reproduction

This is a structured workflow for reproducing deep learning research repos with an audit trail. It walks through README-first target selection, environment setup, execution, and writes standardized reports in a repro_outputs bundle documenting commands, deviations, and scientific comparability. The workflow enforces conservative patching: try command-line fixes before code changes, pause before training runs, and record what breaks README fidelity. It's built for minimal trustworthy runs, not score chasing or open-ended experimentation. The appeal is the discipline: you get evidence logs, a scientific changelog, and machine-readable state instead of silent protocol drift. If you need to verify a paper claim or hand off reproduction context, the structure pays off.

skills.sh installs: 550
GitHub stars: 506
last updated: Jul 14, 2026

View on GitHub

Install

$ npx -y skills add lllllllama/ai-paper-reproduction-skill --skill ai-research-reproduction --agent claude-code

Installs into .claude/skills of the current project.

Files

ai-research-reproduction

Purpose

Guide README-first deep learning reproduction toward the smallest trustworthy run with auditable evidence. Preserve documented meaning; record assumptions, deviations and blockers instead of changing semantics to manufacture success. Load specialized references only for a concrete uncertainty.