prompt-library

📝 プロンプトライブラリ

awesome-chatgpt-prompts とコミュニティのベストプラクティスに着想を得た、実戦で検証済みのプロンプトを集めた包括的なコレクションです。

このスキルを使うタイミング

ユーザーが次のような場合にこのスキルを使用します:

• すぐに使えるプロンプトテンプレートが必要な場合
• ロールベースのプロンプト（「Xとして振る舞って」）を求めている場合
• プロンプトの例やヒントを求めている場合
• タスク固有のプロンプトパターンが必要な場合
• プロンプトの書き方を改善したい場合

プロンプトのカテゴリ

🎭 ロールベースのプロンプト

エキスパート開発者

Act as an expert software developer with 15+ years of experience. You specialize in clean code, SOLID principles, and pragmatic architecture. When reviewing code:
1. Identify bugs and potential issues
2. Suggest performance improvements
3. Recommend better patterns
4. Explain your reasoning clearly
Always prioritize readability and maintainability over cleverness.

コードレビュアー

Act as a senior code reviewer. Your role is to:
1. Check for bugs, edge cases, and error handling
2. Evaluate code structure and organization
3. Assess naming conventions and readability
4. Identify potential security issues
5. Suggest improvements with specific examples

Format your review as:
🔴 Critical Issues (must fix)
🟡 Suggestions (should consider)
🟢 Praise (what's done well)

テクニカルライター

Act as a technical documentation expert. Transform complex technical concepts into clear, accessible documentation. Follow these principles:
- Use simple language, avoid jargon
- Include practical examples
- Structure with clear headings
- Add code snippets where helpful
- Consider the reader's experience level

システムアーキテクト

Act as a senior system architect designing for scale. Consider:
- Scalability (horizontal and vertical)
- Reliability (fault tolerance, redundancy)
- Maintainability (modularity, clear boundaries)
- Performance (latency, throughput)
- Cost efficiency

Provide architecture decisions with trade-off analysis.

🛠️ タスク固有のプロンプト

このコードをデバッグする

Debug the following code. Your analysis should include:

1. **Problem Identification**: What exactly is failing?
2. **Root Cause**: Why is it failing?
3. **Fix**: Provide corrected code
4. **Prevention**: How to prevent similar bugs

Show your debugging thought process step by step.

5歳児にもわかるように説明する（ELI5）

Explain [CONCEPT] as if I'm 5 years old. Use:
- Simple everyday analogies
- No technical jargon
- Short sentences
- Relatable examples from daily life
- A fun, engaging tone

コードのリファクタリング

Refactor this code following these priorities:
1. Readability first
2. Remove duplication (DRY)
3. Single responsibility per function
4. Meaningful names
5. Add comments only where necessary

Show before/after with explanation of changes.

テストを書く

Write comprehensive tests for this code:
1. Happy path scenarios
2. Edge cases
3. Error conditions
4. Boundary values

Use [FRAMEWORK] testing conventions. Include:
- Descriptive test names
- Arrange-Act-Assert pattern
- Mocking where appropriate

API ドキュメント

Generate API documentation for this endpoint including:
- Endpoint URL and method
- Request parameters (path, query, body)
- Request/response examples
- Error codes and meanings
- Authentication requirements
- Rate limits if applicable

Format as OpenAPI/Swagger or Markdown.

📊 分析プロンプト

コードの複雑度分析

Analyze the complexity of this codebase:

1. **Cyclomatic Complexity**: Identify complex functions
2. **Coupling**: Find tightly coupled components
3. **Cohesion**: Assess module cohesion
4. **Dependencies**: Map critical dependencies
5. **Technical Debt**: Highlight areas needing refactoring

Rate each area and provide actionable recommendations.

パフォーマンス分析

Analyze this code for performance issues:

1. **Time Complexity**: Big O analysis
2. **Space Complexity**: Memory usage patterns
3. **I/O Bottlenecks**: Database, network, disk
4. **Algorithmic Issues**: Inefficient patterns
5. **Quick Wins**: Easy optimizations

Prioritize findings by impact.

セキュリティレビュー

Perform a security review of this code:

1. **Input Validation**: Check all inputs
2. **Authentication/Authorization**: Access control
3. **Data Protection**: Sensitive data handling
4. **Injection Vulnerabilities**: SQL, XSS, etc.
5. **Dependencies**: Known vulnerabilities

Classify issues by severity (Critical/High/Medium/Low).

🎨 クリエイティブなプロンプト

機能のブレインストーミング

Brainstorm features for [PRODUCT]:

For each feature, provide:
- Name and one-line description
- User value proposition
- Implementation complexity (Low/Med/High)
- Dependencies on other features

Generate 10 ideas, then rank top 3 by impact/effort ratio.

名前ジェネレーター

Generate names for [PROJECT/FEATURE]:

Provide 10 options in these categories:
- Descriptive (what it does)
- Evocative (how it feels)
- Acronyms (memorable abbreviations)
- Metaphorical (analogies)

For each, explain the reasoning and check domain availability patterns.

🔄 変換プロンプト

コードの移行

Migrate this code from [SOURCE] to [TARGET]:

1. Identify equivalent constructs
2. Handle incompatible features
3. Preserve functionality exactly
4. Follow target language idioms
5. Add necessary dependencies

Show the migration step by step with explanations.

フォーマットの変換

Convert this [SOURCE_FORMAT] to [TARGET_FORMAT]:

Requirements:
- Preserve all data
- Use idiomatic target format
- Handle edge cases
- Validate the output
- Provide sample verification

プロンプトエンジニアリングの手法

Chain of Thought（CoT）

Let's solve this step by step:
1. First, I'll understand the problem
2. Then, I'll identify the key components
3. Next, I'll work through the logic
4. Finally, I'll verify the solution

[Your question here]

Few-Shot Learning

Here are some examples of the task:

Example 1:
Input: [example input 1]
Output: [example output 1]

Example 2:
Input: [example input 2]
Output: [example output 2]

Now complete this:
Input: [actual input]
Output:

ペルソナパターン

You are [PERSONA] with [TRAITS].
Your communication style is [STYLE].
You prioritize [VALUES].

When responding:
- [Behavior 1]
- [Behavior 2]
- [Behavior 3]

構造化された出力

Respond in the following JSON format:
{
  "analysis": "your analysis here",
  "recommendations": ["rec1", "rec2"],
  "confidence": 0.0-1.0,
  "caveats": ["caveat1"]
}

プロンプト改善チェックリスト

プロンプトを作成する際は、次の点を確認してください:

• 明確な目的: 何を正確に求めているか？
• コンテキストの提供: 背景情報を含めていますか？
• フォーマットの指定: 出力をどのように構成すべきか？
• 例の提示: 参考となる例はありますか？
• 制約の定義: 制限や要件はありますか？
• 成功基準: 良い出力をどのように評価しますか？

リソース

• awesome-chatgpt-prompts
• prompts.chat
• Learn Prompting

---

💡 ヒント: 優れたプロンプトは、具体的で、コンテキストを提供し、望ましい出力の例を含んでいます。