What is 'Evals-First' assessment design?

It is a workflow where success criteria are defined based on technical specifications BEFORE questions are written, ensuring assessments accurately measure intended learning outcomes.

Does it support AI-specific coding skills?

Yes, it includes specialized assessment types to measure how well students learn from, teach, and collaborate with AI as a co-worker.

Can it help with grading open-ended code?

Yes, it generates detailed analytic rubrics covering correctness, code quality, efficiency, error handling, and style to ensure consistent and fair grading.

How does this skill ensure assessment quality?

The skill uses a validation script to analyze the cognitive distribution of your assessment, ensuring at least 60% of the questions test application, analysis, or creation rather than simple recall.

AI-Native Assessment Builder

Name: AI-Native Assessment Builder
Author: aiskillstore

byaiskillstore

•

187

•

学習とドキュメンテーション

Generates comprehensive, pedagogically-sound technical assessments and rubrics aligned with Bloom's Taxonomy and AI-native learning patterns.

The AI-Native Assessment Builder helps educators and technical trainers create high-quality, balanced assessments that measure deep conceptual understanding rather than simple memorization. By implementing an 'evals-first' workflow, it ensures that every question maps directly to predefined learning objectives and success criteria. The skill supports a wide range of question formats—from diagnostic MCQs with meaningful distractors to complex debugging challenges and project-based tasks—while maintaining a cognitive distribution where at least 60% of questions test application or higher-order thinking. It uniquely integrates the AI Three-Role Framework to evaluate how effectively students collaborate with AI tools in modern development environments.

主な機能

01Bloom's Taxonomy alignment ensuring 60%+ non-recall cognitive distribution

02Diagnostic MCQ generation with distractors based on common misconceptions

03187 GitHub stars

04Evals-first design workflow mapping questions to specific success criteria

05AI-Native co-learning assessments to evaluate human-AI partnership skills

06Comprehensive analytic rubrics for code-writing and open-ended projects

ユースケース

01Auditing existing curriculum to ensure balanced cognitive depth and CEFR alignment

02Designing midterm or final exams for computer science and programming courses

03Creating formative quizzes for technical documentation, tutorials, and bootcamps

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add aiskillstore/marketplace 92bilal26

For use in Claude.ai and ChatGPT

Download Skill