| name | eli5 |
| description | Re-explain the agent's most recent substantive answer in plain, accessible, jargon-free English. Use when the user types /eli5, says 'explain that simply', 'ELI5 that', 'in plain English', or 're-explain your last answer'. This is a re-render of existing content — it does not introduce new analysis, new options, or a different recommendation. |
| version | 1.0.0 |
| author | Hermes Agent (ported from avidrucker/claude-config eli5 skill design spec) |
| license | MIT |
| platforms | ["linux","macos","windows"] |
| metadata | {"hermes":{"tags":["eli5","explain","plain-language","rephrase","productivity"],"related_skills":[]}} |
ELI5 — Re-explain in Plain English
Re-explain the agent's most recent substantive answer in plain, accessible, jargon-free English. This is a re-render, not a re-do — do not introduce new analysis, new options, or a different recommendation. Only re-explain what is already in the conversation.
Invocation and targeting
/eli5 → re-explain the whole of the agent's most recent substantive answer.
/eli5 <focus> → narrow to a specific piece of it (e.g. /eli5 the register allocation part). The focus may point at an earlier answer if that is clearly what the user means.
Style
- Literal and plain. Short sentences, everyday words, no jargon.
- When a technical term is genuinely unavoidable, define it in one clause.
- No forced analogies or metaphors — explain the real thing directly.
- Keep full technical accuracy underneath the plain words. "ELI5" is the spirit, not literal toddler-speak.
Output structure (adaptive)
Open with a one-sentence plain-English TL;DR, then the body.
- Decision-shaped original answer (tradeoffs / a fork / "should I do A or B") → hit the three beats:
- What the problem is — restated plainly.
- The options — each in plain terms.
- What I recommend and why — the recommendation plus its reasoning.
- Not decision-shaped original answer (explains how something works, walks through a bug, etc.) → fall back to a plain-English walkthrough of whatever the answer actually was. No empty "Options: N/A" sections.
Edge cases
- No prior substantive answer yet — say so, ask what to explain.
- Focus arg matches nothing in the conversation — ask for clarification rather than inventing content.
Out of scope (YAGNI)
- No script, no config, no per-project gating.
- Does not re-run analysis or change the recommendation.
- Not literal toddler-speak.
Source reference
Design spec: docs/superpowers/specs/2026-06-16-eli5-skill-design.md (avidrucker/claude-config).