repo-intake-and-plan

To extract documented commands from a README for AI repository reproduction, this Skill scans project files and README-driven signals to inventory commands and return a structured reproduction plan. It identifies explicit execution instructions for inference, evaluation, and training scenarios.

Classifying repo commands into inference, evaluation, or training tasks is handled by scanning README evidence and config files. The Skill applies conservative classification discipline, relying strictly on explicit documentation to categorize potential execution candidates.

Generating a minimal reproducible plan for an AI repository involves mapping documented commands and proposing the smallest trustworthy target. The Skill outputs an auditable reproduction plan based on README workflows and project scripts suitable for orchestration.

Yes, you can use this to scan project files for reproducible work without running code. It safely reads README files and project configurations to extract static command signals and generate a reproduction plan without executing the actual repository scripts.

Yes, the repository reproduction planner requires explicit README evidence for command classification. It enforces conservative classifications by relying strictly on documented signals in README files and common project workflows rather than inferring commands from source code.