Guide · Agent skill
How to use the Wait What re-pitch agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The wait-what skill is a small instruction file for an AI coding agent. When a reply from the agent is too dense to follow, you ask it to restate its last message with more context, in short controlled English and with the project's own terms. This guide explains what the skill does, how to install it, how to ask for it, and where it does not help.
Wait What Re-pitchmattpocock/skills · MIT Download the skill Quick start
- 1Download the ZIP. Download the wait-what skill ZIP from the page where you found this guide.
- 2Choose where to install. Use ~/.claude/skills/ for personal use, .claude/skills/ inside a project, or the Claude apps under Settings → Capabilities → Skills.
- 3Unzip the skill. Unzip the downloaded file into the folder you chose. If you use the Claude apps, upload the skill there instead.
- 4Check the repo has a glossary. Confirm the repository you are working in contains a GLOSSARY.md file, and GLOSSARY-MAP.md if it has several glossaries.
- 5Invoke the skill after a confusing reply. When the agent's last message does not make sense, ask it deliberately to re-pitch that message with a little context, in simple English, using the glossary.
What does the wait-what skill make the agent do?
The skill re-explains the agent's most recent message. Its instructions ask the agent to give a little bit of context, to write in ASD-STE100 Simplified Technical English, and to use the ubiquitous language defined in a GLOSSARY.md file.
If the repository has more than one glossary, the skill tells the agent to follow GLOSSARY-MAP.md to find the right one. The skill only changes how the explanation is delivered. It does not add tools or access to anything.
- Adds background context to the point that was unclear
- Uses short, controlled Simplified Technical English (ASD-STE100)
- Uses the terms from GLOSSARY.md, or from the glossary GLOSSARY-MAP.md points to
How do you install the wait-what skill?
Download the skill ZIP from the page where you found this guide. Then choose one of the two install options below, depending on how you use the agent.
For a personal setup, unzip the file into ~/.claude/skills/. For a single project, unzip it into the .claude/skills/ folder inside that project. You can also upload the skill in the Claude apps under Settings → Capabilities → Skills.
- Personal: unzip into ~/.claude/skills/
- Single project: unzip into .claude/skills/ in the project
- Claude apps: Settings → Capabilities → Skills, then upload the skill
How do you trigger the skill with a natural request?
This skill is set so the model does not invoke it on its own. You need to call it deliberately when a reply has lost you. Do not expect it to start automatically when you are confused.
The skill's own wording is a good model for what to type. You can say that you do not understand where the agent has got to, and ask it to re-pitch that message with some context, in simple terms, using the project's glossary.
- Explain the gap: say you are lost at the end of the last reply
- Ask for a re-pitch: ask it to restate the message with a little context
- Ask for the project terms: ask it to use the glossary in the repo
What does a good example look like?
Suppose the agent gives a long answer about how two modules interact and uses several domain words you do not know. You reply with a short request to re-pitch that message with context, in simple technical English, using GLOSSARY.md.
The agent then restates the point. The result should use the same words your project uses, which makes it easier to check against your own documentation.
What does the skill need to work?
The skill needs a repository that contains a GLOSSARY.md file. GLOSSARY-MAP.md is optional and is only used when the repo has several glossaries.
The skill runs no scripts and uses no network access. Its instructions are plain text, so there is nothing to execute on your machine.
The terminology step does not apply without a GLOSSARY.md. The skill gives no fallback for that case, so in a repo without one you will get the simplified wording but not the project vocabulary.
- Required: a repository containing GLOSSARY.md, for the terminology step
- Optional: GLOSSARY-MAP.md, if the repo has more than one glossary
- Runs scripts: no
- Uses network: no
What are the safety notes for this skill?
Because the skill runs no scripts and uses no network, it does not send data elsewhere or change files. It changes only how the agent words its explanation.
It also does not fetch new information or take actions. If you need the agent to look something up or edit code, you need a different request.
- No scripts are run
- No network use
- No new tools or access are added
What tips help the re-pitch work better?
Ask soon after the confusing reply, while the context is still in the chat. Keep the request focused on the one message that did not land.
Keep your GLOSSARY.md up to date. The re-pitch can only use the terms the file defines, so clear definitions give clearer explanations.
- Re-pitch right after the message that confused you
- Ask about one point at a time
- Keep GLOSSARY.md clear and current
What people use it for
- Understanding a dense technical explanation from an agent about an unfamiliar part of the codebase
- Working in a repo with a domain glossary and wanting replies that use those exact terms
- Reading agent replies in a second language and wanting shorter, controlled wording
- Checking that the agent's explanation matches the terminology in your project's documentation
Frequently asked questions
Is the wait-what skill free to use?
The skill is published under the MIT licence. The source does not mention pricing, so check the page where you downloaded it for any terms that apply to that copy.
What do I need to use the wait-what skill?
You need a repository that contains a GLOSSARY.md file. GLOSSARY-MAP.md is optional and only matters when the repo has more than one glossary. The skill runs no scripts and uses no network.
Where do I install the wait-what skill?
Unzip the downloaded ZIP into ~/.claude/skills/ for personal use, or into .claude/skills/ inside a project. You can also upload it in the Claude apps under Settings → Capabilities → Skills.
Why doesn't the skill run automatically when I am confused?
The skill is set so the model does not invoke it on its own. You need to call it deliberately after a reply that did not make sense to you.
What happens if my repo has no GLOSSARY.md?
The terminology part of the skill will not apply, and the skill gives no fallback for that case. You will still get the simpler wording, but not the project's own vocabulary.
Does the wait-what skill give me new information or change my code?
No. It only re-explains the agent's last message. It does not fetch or change anything, so use a different request when you need new information or edits.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use wait-what if you often get lost in dense agent replies, work in a repo with a GLOSSARY.md, or prefer shorter controlled English. Pick something else if you need the agent to fetch information or take actions, or if your repo has no glossary, since the terminology step will not apply.
Download the skill