Run a short reproduction and get auditable evidence.
research reproduction
Runs a selected short reproduction command and produces standardized, auditable evidence with patch and comparability reporting when relevant.
When to use it
Use after selecting and preparing a reproduction target, when a short non-training command needs execution evidence and normalized outputs.
Give it a selected reproduction goal, runnable command, and environment and asset assumptions for a project; it runs the command and writes normalized evidence into the project.
What you provide
No additional actions listed in the analysis.
Python 3 is required to run the bundled command runner and output writer.
The skill requires ai-research-reproduction because its bundled shared runtime and writer are loaded from that skill when not available centrally.