prompt-enhancer

name	prompt-enhancer
description	Transforms vague or poorly structured prompts into clear, effective, and well-optimized instructions for AI models. Use when users need help improving their prompts, want better AI responses, or request prompt optimization for any task including coding, writing, analysis, or creative work.

Prompt Enhancer Skill

This skill transforms basic, vague, or poorly structured prompts into clear, specific, and highly effective instructions that produce superior AI outputs. Use this whenever someone needs to improve their prompt quality or wants to get better results from AI models.

Core Principles of Effective Prompts

1. Clarity & Specificity

• Remove ambiguity and vague language
• Define exact requirements and constraints
• Specify format, length, and structure expectations
• Clarify the intended audience and purpose

2. Context Provision

• Include relevant background information
• Define the domain and subject matter
• Specify the user's skill level and knowledge
• Provide examples when helpful

3. Structure & Organization

• Break complex requests into clear steps
• Use numbered lists for sequential tasks
• Use bullet points for parallel requirements
• Organize information hierarchically

4. Role & Persona Definition

• Specify the AI's role or expertise level
• Define the tone and style (formal, casual, technical, etc.)
• Set constraints on the response format
• Clarify the relationship to the user

5. Output Specification

• Define expected format (markdown, JSON, code, etc.)
• Specify length requirements
• Request specific sections or components
• Define success criteria

Enhancement Process

Step 1: Analyze the Original Prompt

Identify issues:

• Vague or unclear requirements
• Missing context or constraints
• Ambiguous language
• Lack of structure
• Missing output specifications
• Insufficient examples

Step 2: Extract Intent

Determine:

• What is the user really trying to achieve?
• What is the expected output?
• Who is the intended audience?
• What level of detail is needed?
• What constraints exist?

Step 3: Enhance Systematically

Apply these improvements:

Add Missing Elements:

• Context and background
• Specific requirements
• Output format specifications
• Examples (when helpful)
• Constraints and limitations

Restructure for Clarity:

• Break into logical sections
• Use clear headings
• Number sequential steps
• Organize parallel requirements

Specify Expectations:

• Tone and style
• Length and format
• Level of detail
• Success criteria

Remove Ambiguity:

• Replace vague terms with specific ones
• Clarify pronouns and references
• Define technical terms if needed
• Specify quantities and measurements

Step 4: Optimize for Model Performance

• Use clear imperative language
• Provide reasoning steps for complex tasks
• Include few-shot examples for format
• Add verification steps for accuracy
• Use delimiters for multi-part prompts

Enhancement Patterns

Pattern 1: The Vague Question

Before: "Tell me about marketing" After:

Act as a digital marketing consultant with 10+ years of experience.

Provide a comprehensive overview of modern digital marketing strategies for a small e-commerce business in 2025.

Include:
1. Top 5 most effective channels (with specific platforms)
2. Budget allocation recommendations (percentage breakdown)
3. Key metrics to track for each channel
4. Common pitfalls to avoid
5. Quick win strategies for the first 90 days

Format: Use clear headings, bullet points for key takeaways, and include 2-3 specific examples for each channel.

Target audience: Business owner with limited marketing knowledge but strong product-market fit.

Length: Approximately 800-1000 words.

Pattern 2: The Unclear Code Request

Before: "Write a function to process data" After:

Write a Python function that processes user registration data with the following specifications:

**Requirements:**
- Function name: `validate_user_registration`
- Input: Dictionary with keys: username, email, password, age
- Output: Tuple of (is_valid: bool, errors: list)

**Validation Rules:**
1. Username: 3-20 characters, alphanumeric + underscore only
2. Email: Valid email format using regex
3. Password: Minimum 8 characters, must contain uppercase, lowercase, number, and special character
4. Age: Integer between 13 and 120

**Error Handling:**
- Return specific error messages for each validation failure
- Multiple errors should all be returned in the list

**Code Style:**
- Use type hints
- Include docstring with examples
- Follow PEP 8 conventions
- Add inline comments for complex regex

**Example Usage:**
```python
result = validate_user_registration({
    "username": "john_doe",
    "email": "[e-mail]",
    "password": "SecureP@ss123",
    "age": 25
})
# Returns: (True, [])

Please include 2-3 test cases demonstrating various error conditions.

### Pattern 3: The Unfocused Analysis Request
**Before:** "Analyze this business idea"
**After:**

Perform a structured business analysis of the following idea:

Business Idea: [Insert specific idea here]

Analysis Framework:

1. Market Opportunity (200 words)

   • Target market size and demographics
   • Current market gaps this addresses
   • Market trends supporting/threatening this idea

2. Competitive Landscape (150 words)

   • Direct competitors (name 3-5)
   • Competitive advantages/disadvantages
   • Differentiation strategy

3. Revenue Model (150 words)

   • Primary revenue streams
   • Pricing strategy considerations
   • Unit economics overview

4. Key Risks (100 words)

   • Top 3 risks with mitigation strategies

5. Go-to-Market Strategy (150 words)

   • Customer acquisition channels
   • Initial target segment
   • Minimum viable product scope

6. Next Steps (100 words)

   • Immediate actions to validate
   • Key metrics to track early
   • Timeline for initial validation

Output Format: Use clear headings, bullet points for key insights, and a executive summary at the top (100 words). Include specific numbers and examples where possible.

Perspective: Analyze from the viewpoint of a seed-stage venture capitalist evaluating investment potential.

### Pattern 4: The Creative Writing Request
**Before:** "Write a story"
**After:**

Write a short science fiction story with the following specifications:

Genre & Tone:

• Hard science fiction with emphasis on realistic technology
• Tone: Thoughtful and introspective, with subtle tension
• No explicit violence or romance

Setting:

• Mars colony in the year 2095
• Small research outpost (6 people)
• Focus on psychological isolation

Plot Requirements:

• Protagonist: Communications engineer discovering an anomaly
• Central conflict: Trust vs. paranoia when unexplained events occur
• Resolution: Ambiguous ending that respects reader intelligence

Technical Constraints:

• Length: 1,500-2,000 words
• Use third-person limited perspective
• Include at least 3 scenes with clear transitions
• Realistic dialogue (no exposition dumps)

Style Guidelines:

• Show, don't tell
• Use sensory details to establish atmosphere
• Ground futuristic elements in current technology
• Avoid clichés like "the red planet" or "brave explorers"

Opening Hook: Start with a specific, intriguing moment in action - not backstory or setting description.

Themes to Explore: Human adaptability, the cost of isolation, the nature of reality

## Prompt Templates by Use Case

### Coding & Development

[Role: Software engineer/architect with X years experience]

Task: [Specific coding task]

Requirements:

• Language/Framework: [Specify]
• Input/Output: [Define clearly]
• Constraints: [Performance, memory, etc.]
• Edge Cases: [List expected edge cases]

Code Quality:

• Style guide: [Specify]
• Documentation: [Level of detail]
• Error handling: [Requirements]
• Testing: [Unit tests needed?]

Context: [Relevant background about the system/project]

Success Criteria: [How to verify the solution works]

### Content Writing

[Role: Content writer/copywriter with expertise in X]

Task: Write [content type] about [topic]

Audience:

• Primary: [Define specifically]
• Knowledge level: [Beginner/Intermediate/Expert]
• Pain points: [What problems they have]

Objectives:

1. [Primary goal]
2. [Secondary goal]
3. [Tertiary goal]

Tone & Style:

• Voice: [Professional/Casual/Technical/etc.]
• Perspective: [First/Second/Third person]
• Style: [Descriptive/Persuasive/Educational/etc.]

Structure:

• Length: [Specific word count or range]
• Format: [Blog post/Email/Landing page/etc.]
• Sections: [Required headings/components]

SEO/Marketing Requirements:

• Keywords: [If applicable]
• Call-to-action: [If applicable]
• Links/References: [Requirements]

Examples: [Provide 1-2 examples of similar successful content]

### Data Analysis

Task: Analyze [dataset/topic] to [specific objective]

Data Context:

• Source: [Where data comes from]
• Volume: [Size/scope]
• Time period: [Relevant timeframe]
• Known issues: [Data quality concerns]

Analysis Requirements:

1. [Specific question 1]
2. [Specific question 2]
3. [Specific question 3]

Methodology:

• Statistical methods: [Specify if known]
• Visualization types: [What charts/graphs]
• Tools/Libraries: [Python/R/SQL/etc.]

Output Format:

• Executive summary: [Length]
• Detailed findings: [Structure]
• Visualizations: [How many, what types]
• Recommendations: [Format]

Success Metrics: [How to measure if analysis is valuable]

### Research & Information Gathering

Research objective: [Specific research question]

Scope:

• Topic boundaries: [What to include/exclude]
• Time period: [If relevant]
• Geographic focus: [If relevant]
• Sources: [Academic/Industry/News/etc.]

Research Questions:

1. [Primary question]
2. [Supporting question 1]
3. [Supporting question 2]

Output Requirements:

• Format: [Report/Summary/Presentation/etc.]
• Length: [Specific or range]
• Citations: [Style and frequency]
• Sections: [Required components]

Quality Standards:

• Source credibility: [Requirements]
• Recency: [How recent must sources be]
• Depth: [Surface level vs. comprehensive]

Perspective: [Neutral/Critical/Supportive analysis]

## Advanced Techniques

### Technique 1: Chain-of-Thought Prompting
Add reasoning steps for complex tasks:

Before providing your answer:

1. Break down the problem into sub-components
2. Analyze each component separately
3. Identify potential issues or edge cases
4. Synthesize findings into a final recommendation

Show your reasoning for each step.

### Technique 2: Few-Shot Learning
Provide examples to establish pattern:

Generate product descriptions following this pattern:

Example 1: Product: Wireless Earbuds Description: "Experience freedom in every beat. These ultra-light wireless earbuds deliver crystal-clear sound for up to 8 hours, with a sleek charging case that fits in your pocket. Touch controls and water resistance make them perfect for workouts, commutes, or just zoning out."

Example 2: Product: Smart Water Bottle Description: "Hydration made intelligent. This app-connected bottle tracks your water intake, glows to remind you to drink, and keeps beverages cold for 24 hours. Whether you're crushing gym goals or desk marathons, staying hydrated has never been this easy."

Now generate a description for: [Your product]

### Technique 3: Constraint-Based Enhancement
Add specific limitations to improve output:

Constraints:

• Do not use: [Forbidden words/approaches]
• Must include: [Required elements]
• Avoid: [Common pitfalls]
• Prioritize: [Key principles]

### Technique 4: Iterative Refinement Structure

Task: [Main task]

First Pass: [Quick initial approach] Review: [What to check for] Refinement: [How to improve] Final Output: [Polish and deliver]

After each step, briefly explain your reasoning.

### Technique 5: Perspective Shifting

Analyze this from three perspectives:

1. [Perspective 1]: [Specific lens]
2. [Perspective 2]: [Different angle]
3. [Perspective 3]: [Contrasting view]

Then synthesize into a balanced recommendation.

## Common Enhancement Patterns

### From Vague to Specific
- "Make it better" → "Improve readability by adding section headers, reducing paragraph length to 3-4 sentences, and using bullet points for lists"
- "Add features" → "Add these specific features: user authentication, data export to CSV, and real-time notifications"
- "Optimize this" → "Optimize for: page load time under 2 seconds, mobile responsiveness, and SEO score above 90"

### From Broad to Focused
- "Marketing strategy" → "Instagram content strategy for B2C SaaS targeting 25-35 year old professionals"
- "Financial analysis" → "Break-even analysis for the next 18 months assuming 15% monthly user growth"
- "Code review" → "Security audit focusing on authentication, input validation, and data encryption"

### From Implicit to Explicit
- Assume role → "Act as a senior DevOps engineer with Kubernetes expertise"
- Assume format → "Provide output as a markdown table with columns: Feature, Priority, Effort, Impact"
- Assume audience → "Explain for a non-technical CEO who needs to make a budget decision"

## Quality Checklist

Before delivering an enhanced prompt, verify:

- [ ] Clear and specific task definition
- [ ] Relevant context provided
- [ ] Expected output format specified
- [ ] Success criteria defined
- [ ] Appropriate role/perspective set
- [ ] Constraints and limitations noted
- [ ] Examples included (if helpful)
- [ ] Ambiguous language removed
- [ ] Logical structure and flow
- [ ] Appropriate level of detail

## Enhancement Output Format

When enhancing a prompt, structure your response as:

Analysis of Original Prompt

[Brief assessment of issues and gaps]

Enhanced Prompt

[The fully enhanced version]

Key Improvements Made

1. [Improvement 1 and why it helps]
2. [Improvement 2 and why it helps]
3. [Improvement 3 and why it helps]

Expected Benefits

[How this enhanced prompt will produce better results]

Optional Variations

[If applicable, suggest alternative approaches or adjustments]

## Examples of Full Enhancement

### Example 1: Research Enhancement

**Original:** "Research AI ethics"

**Enhanced:**

Conduct a focused research analysis on AI ethics in healthcare applications.

Research Scope:

• Focus: Ethical considerations in AI-powered diagnostic tools
• Time period: Last 3 years (2023-2025)
• Geographic focus: US and EU regulatory perspectives
• Source types: Academic papers, regulatory documents, industry reports

Key Research Questions:

1. What are the primary ethical concerns with AI diagnostics?
2. How do current regulations address these concerns?
3. What gaps exist in current ethical frameworks?
4. What best practices have emerged from early adopters?

Output Requirements:

• Format: Structured research report
• Length: 2,000-2,500 words
• Sections:
  • Executive Summary (200 words)
  • Current State of AI in Healthcare Diagnostics
  • Ethical Framework Analysis
  • Regulatory Landscape
  • Case Studies (2-3 examples)
  • Gaps and Recommendations
  • References (APA format)

Analysis Depth:

• Include specific examples and case studies
• Compare and contrast different regulatory approaches
• Identify emerging trends and future considerations

Perspective: Balanced analysis suitable for hospital administrators evaluating AI adoption.

**Key Improvements:**
1. Narrowed from broad "AI ethics" to specific "AI diagnostics in healthcare"
2. Added clear research questions and structure
3. Specified timeframe, geography, and source types
4. Defined exact output format and length
5. Set appropriate perspective for practical application

### Example 2: Coding Enhancement

**Original:** "Fix the bug in my code"

**Enhanced:**

Debug and fix the authentication issue in the following Node.js/Express code.

Problem Description: Users are successfully logging in but the session expires immediately on the next request, requiring them to log in again.

Environment:

• Node.js: v18.17.0
• Express: 4.18.2
• express-session: 1.17.3
• Database: PostgreSQL 14
• Deployment: Docker containers behind nginx reverse proxy

Current Behavior:

1. User submits login form
2. Authentication succeeds
3. User is redirected to dashboard
4. Dashboard renders correctly
5. Any subsequent request shows "not authenticated"

Expected Behavior: Session should persist for 24 hours or until user logs out

Code to Debug: [Paste your code here]

Debugging Requirements:

1. Identify the root cause of session persistence failure
2. Provide the fixed code with inline comments explaining changes
3. Explain why the issue occurred
4. Suggest additional improvements for session security
5. Include environment variable configurations if needed

Additional Context:

• Cookie settings might be affected by nginx proxy
• Using 'secure' and 'httpOnly' flags
• Redis is available if session store needs upgrading

Output Format:

1. Root cause explanation
2. Fixed code (complete middleware setup)
3. Configuration changes needed
4. Testing steps to verify the fix

**Key Improvements:**
1. Specified exact problem (session expiry) vs vague "bug"
2. Provided environment details for context
3. Described current vs expected behavior
4. Asked for explanation, not just code
5. Mentioned relevant deployment context (nginx proxy)

## Tips for Different AI Models

### For Code Generation Models
- Always specify language and framework versions
- Include type hints or interface requirements
- Request test cases and edge case handling
- Specify code style (PEP 8, Airbnb, etc.)

### For Writing Models
- Define tone precisely (e.g., "conversational but authoritative")
- Specify reading level (Flesch-Kincaid score or grade level)
- Request specific rhetorical devices if needed
- Define document structure explicitly

### For Analysis Models
- Request step-by-step reasoning
- Ask for multiple perspectives
- Specify confidence levels for conclusions
- Request supporting evidence and citations

## Anti-Patterns to Avoid

❌ **Don't:**
- Use multiple contradictory instructions
- Provide excessive examples that confuse the pattern
- Make assumptions about shared context
- Use ambiguous pronouns ("it", "this", "that")
- Mix multiple unrelated tasks in one prompt
- Forget to specify output format
- Leave success criteria implicit

✅ **Do:**
- Give one clear primary task
- Provide just enough examples to show the pattern
- Explicitly state all necessary context
- Use specific nouns and references
- Break complex tasks into separate prompts
- Always specify expected output format
- Define clear success criteria

## Final Notes

- **Adaptation is key**: Adjust enhancement depth based on the original prompt's complexity
- **User intent matters**: Sometimes a vague prompt is intentional - ask for clarification if unsure
- **Iterative refinement**: Enhanced prompts can be refined further based on results
- **Balance detail**: Too much specification can be as harmful as too little
- **Test and iterate**: The best prompts emerge through experimentation

Remember: The goal is to transform user intent into AI-actionable instructions that produce consistently excellent results.