/prompt
Basic prompt perfection for quick clarification and refinement.
Overview
The /prompt command analyzes, clarifies, and perfects any prompt into an unambiguous, executable format using the Phase 0 validation flow.
- Time: ~2 seconds
- Complexity: Simple
- Best for: Quick prompts, single-file tasks, well-defined requests
What It Does
- Detects Language - Slovak or English
- Identifies Type - Task, Question, Bug Fix, Feature, Refactoring
- Checks Completeness - Goal, Context, Scope, Requirements, Constraints, Success Criteria
- Asks Clarifying Questions - If information is missing
- Outputs Perfected Prompt - Structured and unambiguous
Usage
bash
/prompt [your prompt]Examples
bash
# Bug fix
/prompt Fix the login bug in my app
# Feature request
/prompt Add dark mode to settings
# Question
/prompt How do I implement caching?
# Refactoring
/prompt Refactor UserService for better performanceThe Phase 0 Flow
Step 0.1: Initial Analysis
Detects:
- Language (Slovak/English)
- Prompt type (Task/Question/Bug/etc.)
- Core intent
Step 0.2: Completeness Check
Verifies presence of:
- ✅ Goal - What you want to achieve
- ✅ Context - Environment, tech stack, background
- ✅ Scope - Files, components affected
- ✅ Requirements - Specific needs
- ✅ Constraints - Limitations, compatibility
- ✅ Success Criteria - How to verify completion
Step 0.3: Clarification
Asks questions about missing items:
- What's unclear?
- Multiple valid approaches?
- Missing context?
Step 0.4: Correction
Fixes:
- Grammar and spelling
- Structure and clarity
- Preserves technical terms
Step 0.5: Perfect Prompt Output
Structured format:
Goal: [One clear sentence]
Context: [Environment, tech stack, background]
Scope: [Files, components]
Requirements:
1. [Requirement 1]
2. [Requirement 2]
Constraints: [Limitations]
Expected Result: [Success criteria]Step 0.6: Approval Gate
You can:
yoryes- Executenorno- Cancelmodify [changes]- Adjustexplain- Get details
Example Flow
Input
bash
/prompt Fix login bugPhase 0 Analysis
Detected Language: English
Prompt Type: Bug Fix
Completeness Check:
✅ Goal: Fix login bug
❌ Context: Which framework? ASP.NET? React?
❌ Scope: Frontend or backend?
❌ Success Criteria: How to verify?
Questions:
1. What login framework are you using?
2. What specific error occurs?
3. Which layer needs fixing (UI, API, database)?After User Answers
Perfected Prompt:
Goal: Fix null reference exception in ASP.NET Core login API
Context:
- ASP.NET Core 6.0
- SQL Server database
- JWT authentication
Scope:
- File: Controllers/UserController.cs
- Method: Login(), line 47
Requirements:
1. Handle null email input gracefully
2. Add proper error logging
3. Return appropriate HTTP status codes
4. Maintain security best practices
Constraints:
- Must maintain existing JWT token structure
- No breaking changes to API contract
Expected Result:
Login API handles invalid input gracefully without throwing exceptions
Changes Made:
- Clarified framework and version
- Identified specific file and method
- Added security requirement
- Defined success criteriaWhen to Use
Use /prompt when:
✅ Quick clarification needed - Simple tasks ✅ Single file changes - Clear scope ✅ No codebase analysis needed - You have context ✅ Well-defined request - Just needs structuring
Use /prompt-hybrid instead when:
❌ Complex task - Multi-file, patterns ❌ Need codebase analysis - Pattern detection ❌ Feasibility check required - Technical validation ❌ Uncertain scope - Let agent explore
Tips for Best Results
- Be specific - Include what you know
- Answer questions - Phase 0 asks for a reason
- Review carefully - Check the perfected prompt
- Use modifications -
modify [change]to refine - Chain commands - Use
/prompt-technicalafter for implementation
Output Structure
All perfected prompts include:
markdown
Goal: [Clear, one-sentence objective]
Context:
- Environment: [OS, platform]
- Tech Stack: [Languages, frameworks]
- Background: [Relevant history]
Scope:
- Files to modify: [List]
- Files to create: [List]
- Components affected: [List]
Requirements:
1. [Specific requirement]
2. [Another requirement]
...
Constraints:
- [Limitation 1]
- [Limitation 2]
Expected Result:
[Definition of success]