Agent Skills: first-principles

|-

UncategorizedID: elasticdotventures/_b00t_/first-principles

Install this agent skill to your local

pnpm dlx add-skill https://github.com/elasticdotventures/_b00t_/tree/HEAD/skills/first-principles

Skill Files

Browse the full folder contents for first-principles.

Download Skill

Loading file tree…

skills/first-principles/SKILL.md

Skill Metadata

Name
first-principles
Description
|-

What

First-principles reasoning means refusing inherited assumptions and deriving conclusions from base truths. Step 1: state what you believe to be true about the problem. Step 2: for each belief, ask "is this a fact, or did I inherit it from convention?" Step 3: discard the inherited beliefs and rebuild from what remains.

Elon Musk's classic example: "batteries are expensive" → why? → materials → materials are cheap → rethink assembly. In software: "we need a database" → why? → persistent state → what's the simplest persistent state? The key test: "would this design make sense to someone who had never seen this domain before?"

When to Use

Apply first-principles thinking to high-stakes decisions where inherited assumptions may be hiding better solutions. It is expensive — do not use it for every micro-choice. Use it when analogy reasoning ("we do it this way because that's how X does it") would lead to a suboptimal outcome.

How

  1. State every belief you hold about the problem.
  2. Challenge each belief — is it a fact or inherited convention?
  3. Discard inherited beliefs. Rebuild from undeniable base truths.
  4. Verify: would this design make sense to a domain newcomer?
  5. Beware the anti-pattern of defaulting to analogy reasoning instead.
<!-- b00t:map v1 summary: first-principles — decompose to base truths; reject inherited assumptions tags: transferable, reasoning, decompose, assumptions, base-truth tier: ch0nky cmds: b00t learn first-principles complexity: 5 -->