aiperf-llm-ergonomics-review

Ergonomics code review for LLM agents evaluates non-mechanical code quality axes like error messaging, type hints, docstrings, and naming to ensure code communicates clearly to AI agents. It identifies readability issues after mechanical checks pass.

Run an ergonomics review on your current branch to evaluate seven quality axes including error-message informativeness, type-hint descriptiveness, and docstring usefulness. This produces a structured critique with explicit citations and concrete rewrite suggestions to guide branch reviews.

The seven ergonomics axes are error-message informativeness, type-hint descriptiveness, docstring usefulness, naming disambiguation, comment semantic density, convention explicitness, and reference-file exemplariness. These axes ensure consistent agent-facing semantics across documentation and code.

Yes, ergonomics review supports pre-release QA by producing structured, actionable critiques after mechanical checks pass. It evaluates new or modified public code paths to ensure consistent agent-facing conventions across documentation and code before release.

Yes, ergonomics code review checks error messages and type hints. It evaluates error-message informativeness and type-hint descriptiveness alongside docstrings, naming, comments, and conventions to improve agent readability and maintainability in AIPerf code.

Ergonomics code review should not be used as a replacement for mechanical checks. It applies specifically to new or modified public code paths in the AIPerf repository and focuses on non-mechanical ergonomics issues rather than functional code correctness.