huh

Restates pending questions or findings in plain language with options and a recommendation.

Updated Jun 2, 2026
One-click install
npx skills add https://github.com/line7works/tony-skills --skill huh-line7works
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: huh
Source: https://github.com/line7works/tony-skills/tree/main/plugins/huh/skills/huh
Command: npx skills add https://github.com/line7works/tony-skills --skill huh-line7works

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When an AI assistant asks a technical question or reports a confusing finding, non-technical users can get stuck decoding jargon before they can decide. This Skill re-explains the pending question or finding in plain language so the user can answer immediately. ## Core Features & Use Cases - Plain-language restatement: Converts the most recent question, finding, or pasted text into a fixed-format block with a one-sentence summary, clearly named options, and a recommended pick. - Multi-decision handling: Repeats the explanation block for each independent decision when a pasted finding bundles several questions. - No-decision mode: Renders purely informational findings without inventing options, and never proceeds with any option until the user explicitly chooses. - Use Case: An assistant reports a confusing build error with three possible fixes. The user types /huh and receives a jargon-free summary of what is being asked, what each fix means in practice, and which one the assistant recommends. ## Quick Start Type /huh after any confusing question or finding to get a plain-language restatement with options and a recommendation.

Frequently Asked Questions about huh

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get an AI assistant to re-explain a confusing question?

Invoke the /huh command or say "explain what you're asking" after the confusing message. The assistant restates the pending question in plain words, lists each option with its practical meaning, and gives a recommended pick.

What does the /huh command do in Claude Code?

The /huh command re-explains the most recent question, finding, or decision point in plain language. It outputs a fixed block: what is being asked, the available options, and a recommendation, then stops and waits for your answer.

Can /huh handle a finding that contains multiple decisions?

Yes. When a pasted finding bundles several independent decisions, the skill repeats its full explanation block for each one, numbered, with an opening sentence stating how many decisions there are.

Does /huh proceed with the recommended option automatically?

No. The skill always stops after presenting options and never proceeds, even with its own recommendation, until the user explicitly picks an option. The reply then resolves the original question in context.

When should I not use /huh?

Do not use it as a casual reaction like "huh, ok do it" — the skill explicitly excludes bare conversational usage. It only triggers on genuine confusion about a pending question or a pasted finding that needs a decision.