code-standards

To enforce code standards when refactoring PyTorch training logic, the skill requires project reconnaissance, an approved implementation plan, and post-change tests with documentation updates. This ensures existing Python and PyTorch modules remain readable and safe.

The best way to maintain consistent Python and shell scripts is applying naming, typing, docstring, and import standards across code changes. It enforces a review-driven workflow requiring project reconnaissance before editing existing repositories.

Yes, the workflow requires an approved implementation plan for non-trivial work before adding new Python modules. It insists on project-wide reconnaissance first to avoid duplicating existing utilities in mature codebases.

To validate configuration file changes after updating a repository, the skill mandates post-change validation. It ensures changes are backed by unit tests, then checked through simplify and review passes before completion.

You need to perform project reconnaissance before editing shell scripts whenever working within an existing repository. This prevents duplicating existing utilities and maintains consistent code standards across the codebase.