alivirgo State Your Goal Not Just the Command (The XY Problem Solver) (AI Skill)
Overview
The XY Problem is a classic cognitive trap: you want to achieve goal X, but you assume mechanism Y is the solution, so you ask an AI for help implementing Y (e.g., "How do I use a complex regex to extract the last 3 letters of a filename?").
The AI will faithfully answer your literal question (writing a fragile regex), whereas if you had stated your true goal X ("I want to check if a file is an image"), the AI would provide the vastly superior 1-line native solution (filename.endswith(('.png', '.jpg'))).
The Underlying Intent Protocol ensures you always state the ultimate outcome, enabling the AI to recommend simpler, more elegant approaches.
The XY Problem Architecture
┌─────────────────────────────────────────────────────────────┐
│ The Classic XY Problem │
│ │
│ TRUE GOAL (X): "I need to check if uploaded files are PNGs"│
│ │ │
│ ▼ (User assumes wrong path) │
│ ATTEMPTED FIX (Y): "Write a complex regex on file string" │
│ │
│ • Asking for Y only $\rightarrow$ Fragile, 20-line regex │
│ • Stating X + Y $\rightarrow$ "Use `pathlib.suffix` or │
│ MIME validation instead" │
└─────────────────────────────────────────────────────────────┘
The 2-Part Intent Formula
┌───────────────────────────────────────────────────────────────────────────┐
│ 1. THE TASK I AM TRYING TO DO: "I was trying to write a custom cron job..."│
│ 2. THE ULTIMATE GOAL (WHY): "...because I need to auto-delete temporary │
│ uploads older than 24h from AWS S3." │
└───────────────────────────────────────────────────────────────────────────┘
Master Intent-First Prompt Templates
Pattern 1: The "Outcome + Attempted Path" Template
Use for software architecture, devops, and database design:
Here is what I am trying to achieve:
- **Ultimate Goal (X)**: [DESCRIBE THE FINAL OUTCOME / USER EXPERIENCE]
- **My Current Idea (Y)**: [DESCRIBE WHAT YOU WERE PLANNING TO DO]
Questions:
1. Is my current idea (Y) the optimal way to achieve (X)?
2. If yes: provide the implementation.
3. If no: what is the simpler, standard industry pattern to achieve (X)?
Pattern 2: The Business & Operational Intent Directive
Use for workflows, automation, and spreadsheet logic:
I want to [INSERT BUSINESS GOAL: e.g. Automatically notify sales reps when a lead books a demo].
Instead of just answering my literal question, let me know if there is a zero-code or built-in tool that accomplishes this faster.
Real-World Case Study
Scenario: Sizing Cloud Instances
Literal Command (Trapped in Y)
User: "How do I configure a custom Python daemon on an EC2 instance to poll a database every 5 seconds for new customer signups and send a Slack alert?"
❌ AI gives 80 lines of Python daemon code, systemd service units, and connection pooling.
Intent-First Prompt (Solved with X)
User: "Goal: We need real-time Slack alerts when a new customer signs up in our PostgreSQL database. I was thinking of running a polling script on an EC2 server, but what is the simplest, lowest-maintenance way to do this?"
AI Response:
Superior Solution: Do not run a server or polling daemon. Use a native PostgreSQL Database Webhook / Supabase Edge Function or an AWS Lambda function on DynamoDB Streams to push a webhook to Slack directly on row insertion.
- Cost: $0.00 / month (Runs within free tier vs $15/mo EC2 server).
- Maintenance: Zero servers to patch or keep alive.
Summary Best Practices
- Always include the word "Because": Adding "I need to do this because [GOAL]" unlocks the model's architectural reasoning.
- Invite better alternatives: Explicitly prompt: "If there is a better way to solve this, challenge my approach."