Eli5

---
name: eli5
description: "Explains any document, concept or architecture in plain adult language, with analogies and no jargon. Triggers on 'eli5', 'explicame como si tuviera 5', 'hazlo simple', 'en cristiano'. Alias of lii5."
metadata:
  short-description: ELI5 alias (universal naming for lii5)
  triggers: "eli5, ELI5, /eli5, explain like i'm 5, explícame, hazlo simple, dumb it down, en cristiano"
  domain: explanation / accessibility / pedagogy
---

# eli5 — Explain Like I'm 5 (alias of lii5)

This skill is a **universal-naming alias** for `lii5`. They produce the same output style;
they just respond to different invocation names.

| You can say | Skill that loads |
|---|---|
| `/eli5`, `eli5`, `ELI5`, "explain like I'm 5" | This skill |
| `/lii5`, `lii5`, `li5`, "explícame como si tuviera 5" | `lii5` skill |

Both apply the same rules. See `~/.claude/skills/lii5/SKILL.md` for the canonical workflow.

## Why a separate file

The slash-command system (and most LLM trigger discovery) looks up skill files by name.
A user typing `/eli5` won't find `/lii5` unless we maintain a discoverable mapping. Two
small files is cheaper than a redirect layer.

## Style — the canonical rules (mirror of lii5)

When invoked, produce explanations that follow:

### Use adult language, not childish

- ❌ "Imagine you have a toy box…"
- ✅ "Think of it like a filing cabinet at the office…"

The audience is a competent adult who simply doesn't know this specific domain. Treat
them as such.

### Lead with one strong analogy

The first paragraph should be a real-world analogy that captures the **shape** of the
concept. Then unpack the technical details against that scaffold.

### Avoid jargon-stacking

If you have to use a technical term, explain it the first time inline:
- ❌ "It uses a CDP-driven Chromium instance with persistent CDP sessions"
- ✅ "It uses Chrome DevTools Protocol (CDP) — the same wire protocol Chrome's own DevTools uses — to drive a regular Chromium browser. Because the CDP session can outlive a single command, you get a persistent browser instance you can come back to."

### Use tables, arrow flows, and code blocks where they help

A 5-row table beats three paragraphs of prose. An ASCII arrow diagram (`A → B → C`) beats
a sentence describing a flow. Use them when they earn their place.

### Close with a one-sentence summary

After the unpacking, give a single-sentence "TL;DR in plain English" — the version
someone would repeat back at a meeting.

## When to use eli5 vs other skills

| Use eli5 / lii5 | Use something else |
|---|---|
| Concept the user doesn't know yet | Decision needing structure → use the formal doc skills |
| Technical jargon that needs unpacking | Persuasion / negotiation → use voice-and-cadence-consistency |
| "What is X?" / "How does Y work?" | Code review → use code-review skill |
| Bridging a gap before a decision | Implementation steps → use the specific tool skill (postgres, kubernetes, etc.) |

## Anti-patterns (do NOT do these)

- ❌ Open with "Great question!" or "Let me explain this carefully" — corporate-LLM tell
- ❌ Use childish framing ("imagine you're a wizard…")
- ❌ Wall of unbroken prose — break with subheadings, tables, diagrams
- ❌ Over-claim certainty when the underlying tech has tradeoffs — note them
- ❌ Drop a Wikipedia paragraph — use lived analogies

## Lessons Learned

- 2026-05-11: Created as `/eli5` alias because user typed the standard internet shorthand and got "Unknown command. Did you mean /lii5?" The fix is to maintain both invocation paths.