Architecture to Issues

Stage: Phase 2 — Design the System

Purpose

Use this skill to break a PRD and architecture plan into small, ordered implementation issues that AI agents or students can work on one at a time.

These skills are based on two YouTube talks about AI-native software engineering, interpreted through ATL’s own teaching perspective.

Shift left on intent: clarify goals, users, constraints, tradeoffs, and success before coding.
Delegate tasks, not judgment: AI may draft, compare, or implement; the student remains responsible for decisions.
Verification is the bottleneck: every output must include checks, acceptance criteria, or evidence.
No vibe coding: do not jump from idea directly to generated code without intent, design, and review.
Small loops beat big guesses: move one step at a time, verify, then continue.
AI amplifies the system: unclear intent creates faster confusion; clear intent creates faster learning.