grey-haven-prompt-engineering
について
このスキルは、LLMの応答品質を大幅に改善する26の文書化されたプロンプトエンジニアリングの原則とテンプレートを提供します。技術的タスク、創造的タスク、研究タスク向けのアンチパターン、チェックリスト、ガイダンスを含みます。開発者は、プロンプトの作成、AI応答の最適化、またはエージェント指示の設計時にこれを活用すべきです。
クイックインストール
Claude Code
推奨/plugin add https://github.com/greyhaven-ai/claude-code-configgit clone https://github.com/greyhaven-ai/claude-code-config.git ~/.claude/skills/grey-haven-prompt-engineeringこのコマンドをClaude Codeにコピー＆ペーストしてスキルをインストールします
ドキュメント
Prompt Engineering Skill
Master 26 documented principles for crafting effective prompts that get high-quality LLM responses on the first try.
Description
This skill provides comprehensive guidance on prompt engineering principles, patterns, and templates for technical tasks, learning content, creative writing, and research. Improves first-response quality by 400%+.
What's Included
Examples (examples/)
- Technical task prompts - 5 transformations (debugging, implementation, architecture, code review, optimization)
- Learning task prompts - 4 transformations (concept explanation, tutorials, comparisons, skill paths)
- Common fixes - 10 quick patterns for immediate improvement
- Before/after comparisons - Real examples with measured improvements
Reference Guides (reference/)
- 26 principles guide - Complete reference with examples, when to use, impact metrics
- Anti-patterns - 12 common mistakes and how to fix them
- Quick reference - Principle categories and selection matrix
Templates (templates/)
- Technical templates - 5 ready-to-use formats (code, debug, architecture, review, performance)
- Learning templates - 4 educational formats (concept explanation, tutorial, comparison, skill path)
- Creative templates - Writing, brainstorming, design prompts
- Research templates - Analysis, comparison, decision frameworks
Checklists (checklists/)
- 23-point quality checklist - Verification before submission with scoring (20+ = excellent)
- Quick improvement guide - Priority fixes for weak prompts
- Category-specific checklists - Technical, learning, creative, research
Key Principles (Highlights)
Content & Clarity:
- Principle 1: No chat, concise
- Principle 2: Specify audience
- Principle 9: Direct, specific task
- Principle 21: Rich context
- Principle 25: Explicit requirements
Structure:
- Principle 3: Break down complex tasks
- Principle 8: Use delimiters (###Headers###)
- Principle 17: Specify output format
Reasoning:
- Principle 12: Request step-by-step
- Principle 19: Chain-of-thought
- Principle 20: Provide examples
Impact Metrics
| Task Type | Weak Prompt Quality | Strong Prompt Quality | Improvement |
|---|---|---|---|
| Technical (code/debug) | 40% success | 98% success | +145% |
| Learning (tutorials) | 50% completion | 90% completion | +80% |
| Creative (writing) | 45% satisfaction | 85% satisfaction | +89% |
| Research (analysis) | 35% actionable | 90% actionable | +157% |
Use This Skill When
- LLM responses are too general or incorrect
- Need to improve prompt quality before submission
- Training team members on effective prompting
- Documenting prompt patterns for reuse
- Optimizing AI-assisted workflows
Related Agents
prompt-engineer- Automated prompt analysis and improvementdocumentation-alignment-verifier- Ensure prompts match documentation- All other agents - Improved agent effectiveness with better prompts
Quick Start
# Check quality of your prompt
cat checklists/prompt-quality-checklist.md
# View examples for your task type
cat examples/technical-task-prompts.md
cat examples/learning-task-prompts.md
# Use templates
cat templates/technical-prompt-template.md
# Learn all principles
cat reference/prompt-principles-guide.md
RED-GREEN-REFACTOR for Prompts
- RED: Test your current prompt → Likely produces weak results
- GREEN: Apply principles from checklist → Improve quality
- REFACTOR: Refine with templates and examples → Achieve excellence
Skill Version: 1.0 Principles Documented: 26 Success Rate: 90%+ first-response quality with strong prompts Last Updated: 2025-01-15
GitHub リポジトリ
関連スキル
content-collections
メタThis skill provides a production-tested setup for Content Collections, a TypeScript-first tool that transforms Markdown/MDX files into type-safe data collections with Zod validation. Use it when building blogs, documentation sites, or content-heavy Vite + React applications to ensure type safety and automatic content validation. It covers everything from Vite plugin configuration and MDX compilation to deployment optimization and schema validation.
creating-opencode-plugins
メタThis skill provides the structure and API specifications for creating OpenCode plugins that hook into 25+ event types like commands, files, and LSP operations. It offers implementation patterns for JavaScript/TypeScript modules that intercept and extend the AI assistant's lifecycle. Use it when you need to build event-driven plugins for monitoring, custom handling, or extending OpenCode's capabilities.
evaluating-llms-harness
テストThis Claude Skill runs the lm-evaluation-harness to benchmark LLMs across 60+ standardized academic tasks like MMLU and GSM8K. It's designed for developers to compare model quality, track training progress, or report academic results. The tool supports various backends including HuggingFace and vLLM models.
sglang
メタSGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.