Grade And Review

Use this skill to evaluate student submissions against BKT-informed 4-tier criterion-referenced rubrics, eliminate AI grading hallucinations via verifiable line citations and executable verification, diagnose cognitive error categories (including AI-mediated errors), and deliver evidence-based growth-mindset feedback enforcing the Rule of One with Socratic hint ladder.

When to Use

• evaluating student exercises, code repositories, or open-ended design submissions
• grading against BKT-weighted 4-tier analytic rubrics (Beginning, Developing, Proficient, Advanced)
• diagnosing cognitive failure modes (conceptual vs procedural vs boundary vs overload vs AI-mediated)
• authoring evidence-based growth-mindset feedback using "not-yet" framing + metacognitive strategy training
• emitting verified machine-readable assessment contracts under EU AI Act High-Risk AI Governance (Annex III, 3a)

Core Rules