# Explanation Design

> Explanation Design

- Skill: `sethdford/explanation-design` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add sethdford/explanation-design`
- Raw SKILL.md: https://api.skillmd.com/api/skills/sethdford/explanation-design/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: sethdford (https://skillmd.com/u/sethdford)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/sethdford/explanation-design

---

# Explanation Design

Match depth to the listener: right entry point, controlled new concepts, and checks so understanding is real—not performative nodding.

## When to Use
- Onboarding, architecture walkthroughs, teaching protocols, or translating expert knowledge for a mixed audience.

## Workflow
1. **Assess prior knowledge** in one or two questions; do not assume zero or expert by title alone.
2. **Choose entry point** from where they are—not always “start at the beginning.”
3. **Progressive disclosure:** crisp overview first; invite “go deeper” on branches they care about.
4. Teach **one new concept at a time**; link each to a **concrete example** before the abstract definition.
5. **Check understanding** with a small application or paraphrase before advancing.
6. **Feynman test:** if you cannot explain it simply without hand-waving, refine the model or analogy.

## Examples
**Example 1:** Explain caching → start from “repeated expensive work,” show one request trace, then name layers and tradeoffs.

**Example 2:** Expert audience → skip basics, lead with invariants and failure modes they can validate.