스킬 prompt-engineer
🎯

prompt-engineer

안전

使用经过验证的框架优化您的AI提示词

不再为模糊的AI回复而烦恼。这个技能将您的初步想法转化为精确的、经过框架优化的提示词,从任何AI助手那里获得一致的高质量结果。

지원: Claude Codex Code(CC)
📊 71 적절함
1

스킬 ZIP 다운로드

2

Claude에서 업로드

설정 → 기능 → 스킬 → 스킬 업로드로 이동

3

토글을 켜고 사용 시작

테스트해 보기

"prompt-engineer" 사용 중입니다. improve this prompt: help me write Python code

예상 결과:

  • The skill asks for clarification: What type of Python code do you need?
  • After selecting 'Web scraping', outputs a structured prompt with Role (Python developer expert), Task (create web scraping script), Requirements (BeautifulSoup, rate limiting, error handling), and Output format (complete commented code)

"prompt-engineer" 사용 중입니다. create prompt for: product recommendation system with purchase history and market trends

예상 결과:

  • Generates a comprehensive RODES + Chain of Thought prompt with Role (ML engineer), Objective (design recommendation system), Details (data sources, algorithms, metrics), Step-by-step approach, Examples (pseudocode, diagrams), Sense Check validation, and Output format (technical design document)

보안 감사

안전
v1 • 2/24/2026

All 148 static analysis findings are false positives. The detected patterns exist in documentation files (README.md, SKILL.md) showing example usage in markdown code blocks, not executable code. Shell command examples are instructional text for users, path references are markdown relative links, and 'cryptographic' detections are framework names like 'Chain of Thought'. No actual security risks identified.

2
스캔된 파일
914
분석된 줄 수
0
발견 사항
1
총 감사 수
보안 문제를 찾지 못했습니다
감사자: claude

품질 점수

38
아키텍처
100
유지보수성
87
콘텐츠
31
커뮤니티
100
보안
91
사양 준수

만들 수 있는 것

寻求更好代码生成的开发者

将模糊的编码请求转化为精确的技术提示词,指定需求、约束条件和预期输出格式,以获得更准确的AI生成代码

提高AI写作的内容创作者

将基本的写作想法转化为结构化的提示词,包含受众、语气、格式和内容指南,以获得一致的高质量AI生成内容

增强数据分析洞察的分析师

将一般的分析请求转化为有条理的方法论提示词,使用RISE或Chain of Thought等框架进行系统化、彻底的AI驱动分析

提示词模板

이 프롬프트를 사용해 보세요

Quick Prompt Enhancement
Improve this prompt: [paste your rough prompt here]
Complex Task Structuring
Create a detailed prompt for: [describe your complex goal with multiple requirements]
Framework-Specific Request
Transform this into a Chain of Thought prompt: [your problem that needs step-by-step reasoning]
Role-Based Prompt Generation
Create a prompt where the AI acts as a [specific expert role] to help me [specific task]

모범 사례

  • Provide specific context about your goal rather than generic requests for better results
  • Select 'Web scraping' from the multiple-choice options when the skill asks for clarification to get a more targeted prompt
  • Use the refined prompt directly with your AI assistant without modification for consistent results

피하기

  • Do not expect the skill to explain which framework it used unless you explicitly ask
  • Avoid providing overly vague input like 'help me improve' without specifying what to improve
  • Do not mix multiple unrelated tasks in a single prompt optimization request

자주 묻는 질문

Do I need to learn prompting frameworks to use this skill?
No. The skill automatically selects and applies the best framework(s) based on your task. You simply describe what you want, and it handles the framework selection silently.
How many questions will the skill ask me?
Maximum 2-3 questions, and only when critical information is missing. Most prompts are optimized directly without any clarification needed.
Can I use the generated prompts with any AI model?
Yes. The prompts are model-agnostic and work with Claude, ChatGPT, Gemini, or any conversational AI assistant.
What if I want a shorter or longer prompt?
You can request adjustments like 'make it shorter' or 'add more detail' after receiving the optimized prompt, and the skill will refine it accordingly.
Does this skill work in any project context?
Yes. It is a universal skill that works in any terminal context without requiring specific project structure, Obsidian vaults, or external configuration.
What languages are supported?
The skill adapts to your input language. If you provide input in Portuguese, it generates the prompt in Portuguese. Mixed inputs default to English.

개발자 세부 정보

파일 구조

📄 README.md

📄 SKILL.md