Claude Prompt Refiner

Refine prompts to get better results from Claude models by applying Anthropic's recommended patterns.

When to Use

Invoke this skill when you have a task for Claude that:

• Involves multiple steps or files

• Requires specific output formatting

• Needs careful reasoning or analysis

• Would benefit from structured context

Refinement Process

Step 1: Analyze the Draft Prompt

Review the user's prompt for:

• Clear outcome definition

• Sufficient context

• Explicit constraints

• Success criteria

Ask clarifying questions if any of these are missing.

Step 2: Apply Claude-Specific Patterns

Structure with XML tags: Claude responds exceptionally well to XML-style tags for organizing complex prompts:

• <context>

• Background information, codebase state, environment

• <task>

• The specific action to take

• <requirements>

• Must-have criteria

• <constraints>

• Limitations and boundaries

• <examples>

• Sample inputs/outputs if helpful

• <output_format>

• How to structure the response

Ordering matters:

• Context first (what exists)

• Task second (what to do)

• Requirements third (how to do it)

• Examples last (clarifying edge cases)

Step 3: Enhance for Reasoning

For complex tasks, add thinking prompts:

• "Think through the approach before implementing"

• "Consider these edge cases: ..."

• "Explain your reasoning for key decisions"

Step 4: Output the Refined Prompt

Present the improved prompt with:

• Clear section headers

• XML tags where beneficial

• Specific, measurable criteria

• An explanation of what changed and why

Example Transformations

Example 1: Vague Task

Before:

Add caching to the API

After:

<context> The /api/products endpoint currently queries the database on every request. Average response time is 200ms. We use Redis for other caching in the app. </context>

<task> Add Redis caching to the /api/products endpoint to reduce database load. </task>

<requirements>

• Cache TTL of 5 minutes
• Cache invalidation when products are updated
• Graceful fallback to database if Redis is unavailable
• Add cache hit/miss metrics logging </requirements>

<constraints>

• Don't change the response format
• Must pass existing integration tests
• Use our existing Redis connection from src/lib/redis.ts </constraints>

Example 2: Code Review Request

Before:

Review this PR

After:

<context> This PR adds user authentication to our Next.js application. Our stack: Next.js 14, TypeScript, Prisma, PostgreSQL. Security is critical - this handles user sessions and passwords. </context>

<task> Review the changes in this PR for security issues, code quality, and adherence to our patterns. </task>

<requirements> Focus on:

1. Security vulnerabilities (auth bypass, injection, etc.)
2. Error handling and edge cases
3. TypeScript type safety
4. Test coverage for critical paths </requirements>

<output_format> Organize your review as:

Critical Issues (must fix before merge)

Recommendations (should consider)

Minor Suggestions (nice to have)

What Looks Good (positive feedback)

</output_format>

Example 3: Feature Implementation

Before:

Add dark mode

After:

<context> React application using Tailwind CSS for styling. Currently only has light mode. Design tokens are in tailwind.config.js. User preference should persist across sessions. </context>

<task> Implement dark mode toggle with system preference detection and persistence. </task>

<requirements>

• Toggle component in the header
• Detect system preference on first visit
• Persist user choice in localStorage
• Smooth transition between modes
• Update all existing components to support both modes </requirements>

<constraints>

• Use Tailwind's built-in dark mode support
• Don't add new dependencies
• Ensure WCAG AA contrast ratios in both modes </constraints>

<examples> Current light mode colors:

• Background: bg-white
• Text: text-gray-900
• Primary: text-blue-600

Expected dark mode equivalents:

• Background: dark:bg-gray-900
• Text: dark:text-gray-100
• Primary: dark:text-blue-400 </examples>

Tips for Best Results

• Be specific about scope - "the auth module" → "src/auth/session.ts"

• Include file paths when relevant

• Reference existing patterns - "follow the pattern in UserService.ts"

• State what NOT to do - constraints prevent unwanted changes

• Define done - what does success look like?