Content creators need systematic, objective quality assessment to improve educational materials. This skill provides a 6-category weighted rubric with evidence-based scoring to consistently evaluate and improve book chapters and lessons.
スキルZIPをダウンロード
Claudeでアップロード
設定 → 機能 → スキル → スキルをアップロードへ移動
オンにして利用開始
テストする
「content-evaluation-framework」を使用しています。 Evaluate this lesson file: apps/learn-app/docs/chapter-3/lesson-2.md
期待される結果:
- **Overall Verdict**: Good (81.55%)
- **Constitution Compliance**: Pass
- **Technical Accuracy**: Good (80%) - Code works, minor type hint gaps
- **Pedagogical Effectiveness**: Excellent (92%) - Strong scaffolding, great exercises
- **AI-First Teaching**: Needs Work (65%) - AI exercises present but weak guidance
- **Key Improvement**: Strengthen AI exercise guidance section (lines 45-52)
セキュリティ監査
安全Pure prompt-based documentation skill containing only evaluation rubrics, checklists, and templates. No executable code, network access, file system access beyond its own directory, or external command execution capabilities.
品質スコア
作れるもの
Mid-Draft Feedback
Get early quality checkpoints on partial drafts to catch foundational issues before completion.
Completion Assessment
Conduct comprehensive evaluation of completed lessons against publication standards.
Pre-Validation Gate
Verify content meets minimum quality thresholds before entering SDD validation phase.
これらのプロンプトを試す
Use the Content Evaluation Framework to evaluate [file_path]. Check constitution compliance first, then score each category using the rubric. Report overall tier and specific improvement recommendations.
Evaluate the [specific section] section of [file_path] for [Technical Accuracy/Pedagogical Effectiveness/Writing Quality]. Highlight any issues and suggest improvements.
Run a constitution compliance check on [file_path] using the constitution-checklist.md. Report pass/fail status with specific violations if any.
Conduct a full evaluation of [file_path] against publication standards. Calculate weighted score, verify all quality gates pass, and provide actionable next steps for final approval.
ベストプラクティス
- Start with constitution compliance check before weighted evaluation - it is a gate that blocks publication if failed
- Provide specific evidence (quotes, line numbers) for every score rather than subjective judgments
- Prioritize recommendations by impact: critical blocking issues first, then important, then enhancement opportunities
回避
- Skipping constitution compliance check and jumping straight to weighted scoring
- Providing vague feedback like 'improve clarity' without specifying which sentences need work
- Rating content 'Good' when it should be 'Needs Work' to avoid difficult feedback conversations
よくある質問
What AI tools are compatible with this skill?
What is the minimum score for publication?
How long does an evaluation take?
Does this skill modify the content files?
What if constitution compliance fails?
How is this different from general editing?
開発者の詳細
作成者
92Bilal26ライセンス
MIT
リポジトリ
https://github.com/92Bilal26/TaskPilotAI/tree/main/.claude/skills/content-evaluation-framework参照
main
ファイル構成