← ClaudeAtlas

eli5listed

Explain like I'm 5: picture-first HTML, any model. /eli5.
sscodeai/eli5 · ★ 0 · AI & Automation · score 73
Install: claude install-skill sscodeai/eli5
# ELI5 Skill — Big pictures, few words · any model · multi-language · interactive Explain any topic to someone who knows **nothing** about it, as a **single-file HTML** output: large hand-drawn-style SVG figures tell the story, text stays in the background. No dependency on any specific model's Artifacts or proprietary tools — any LLM (Claude, GPT, DeepSeek, Gemini, local models) that can write HTML can follow this spec and produce the same output. Supports **中文 / English / 日本語 / 한국어** with auto-detection or explicit selection, and one-shot generation of the full multi-language set. ## When to Use - User types `/eli5 <topic>` (this skill's slash trigger) - User says "explain X like I'm 5" / "用大白话讲一下 X" / "解释给我妈听" / "子どもに説明する" / "쉽게 설명해줘" - User asks for a "simple visual explanation" / "one-picture科普" of a concept - User asks for "multi-language version" / "中英日" / "multi-language explanation" - When the topic lives in the codebase/session, `read_file` the relevant code first so the explanation is accurate Don't use for: deep technical documentation, teaching that needs precise terminology, plain-text Q&A (no visualization needed). ## Language rules (multi-language core) 1. **Auto-detect**: look at the user's message language + session context. Chinese message → 中文; English → English; 日本語 → 日本語; 한국어 → 한국어. 2. **Explicit request wins**: `/eli5 X in Japanese`、`/eli5 X 用日语`、`/eli5 X を日本語で` → output in the requested language. 3. **One-shot multi-language**: `/eli5 X 多语言`、`/el