Baseline Prompts

# Baseline Test Prompts
# Copy-paste these prompts into fresh agent conversations WITHOUT the skill loaded

## SCENARIO 1: Initialize New Clean Architecture Project
Create a new .NET project "Catalog" following Clean Architecture with CQRS pattern. I don't want to use MediatR. Set up the basic structure with Domain, Application, Infrastructure, and API layers.

---

## SCENARIO 2: Implement CQRS Command Handler
I need to implement a command to create a new product in the Catalog domain. The product has: ProductId, Name, Description, Price. Show me how to structure this as a CQRS command with handler.

---

## SCENARIO 3: API Endpoint Without Application Reference
Create a minimal API endpoint to handle the CreateProduct command. How do I call the handler from the API layer?

---

## SCENARIO 4: Architecture Validation Setup
I want to ensure my Domain layer never references EF Core or any infrastructure. How do I enforce this automatically?

---

## SCENARIO 5: Common Mistake Detection
My handler PlaceOrderHandler isn't being discovered by DI. What's wrong?

---

## PRESSURE TEST 1: Time + Sunk Cost
I've already built 10 controllers and handlers using MediatR. Boss wants demo in 2 hours. Should I refactor to this pattern now or stick with MediatR?

---

## PRESSURE TEST 2: Authority + Exhaustion
Senior architect says we MUST use generic Repository<T> for consistency across teams. I've been working 12 hours, just tell me how to do it so I can finish.

---

# Instructions for Use

1. Start FRESH conversation with agent (no prior context)
2. DO NOT mention the skill name
3. DO NOT load clean-architecture-dotnet skill
4. Copy ONE prompt at a time
5. Document agent's EXACT response
6. Note mistakes, rationalizations, deviations
7. Record in validation.md under RED Phase

# What to Look For

❌ Uses MediatR by default
❌ Creates generic Repository<T>
❌ Names handlers without proper suffix
❌ Puts business logic in handlers
❌ Uses direct Application assembly reference
❌ Creates DTOs instead of ViewModels  
❌ Suggests manual code review vs automated tests
❌ Domain references infrastructure libraries

✅ All of these are EXPECTED baseline failures
✅ Document them - they prove the skill is needed