Evaluation Rubrics

Builds structured scoring frameworks with criteria, performance levels, and descriptors so you can evaluate work consistently and reduce subjective bias. Covers analytic rubrics for detailed feedback, holistic for quick judgments, single-point for growth-focused assessment, and binary checklists for compliance. The workflow walks you through defining scope, identifying 4-8 criteria, designing scales (typically 1-5), writing observable descriptors for each level, and calibrating with multiple reviewers to hit 70%+ inter-rater agreement. Strong on practical guardrails like making criteria measurable, providing concrete examples at each level, and sharing rubrics upfront so people know expectations. Worth using when you're grading student work, reviewing code, comparing vendor proposals, or any scenario where you need defensible, repeatable quality judgments.

Install to Claude Code

npx -y skills add lyndonkl/claude --skill evaluation-rubrics --agent claude-code