// Mar 21, 2026
What It Does
First Principles is an AI skill that deconstructs complex challenges into foundational truths, then reconstructs them as radically better solutions. It applies the same analytical methodology used by Elon Musk at SpaceX and Tesla — stripping away inherited assumptions until only physics, logic, and economics remain.
The skill operates in four stages:
Why I Built It
As a designer, I notice that most design decisions are inherited. We pick the nav pattern everyone else uses. We follow the conventions of whatever product category we're working in. We optimize within constraints we never questioned.
First principles thinking breaks this. It asks: if we ignored everything that exists, what would we actually build? The answer is almost always different — and often better.
I built this skill because the method is powerful but hard to apply rigorously without a structured guide. Most people stop at "question your assumptions" and never get to the reconstruction phase. This skill forces the full process — from deconstruction to a testable MVP.
It's equally useful for product decisions, architecture choices, and design direction. Anywhere you feel stuck inside a box you didn't build, this skill helps you think outside it.
How to Use It
$ step_01
Go to the first-principles-skill GitHub repository and copy the content of SKILL.md. Review the example cases to understand the expected depth of analysis.
$ step_02
Paste SKILL.md as a system prompt or custom instruction in:
$ step_03
Fill in input.yaml with your challenge — describe the problem, the current solution you're considering, and any constraints you believe are fixed. The more context you provide, the deeper the deconstruction.
$ step_04
The AI guides you through each stage in sequence, pushing back on assumptions and refusing to accept inherited conventions as constraints. The output is a reconstructed approach and a concrete MVP to test it.