wait-what Skill — Make AI Explain the Previous Unclear Answer Again

Skill Overview

wait-what is a manually triggered Claude skill. When the previous AI response was not clear, it requires the AI to add context, use ASD-STE100 Simplified Technical English, and explain the same thing again with the standard terminology defined in CONTEXT.md.

Use Cases

1. Taking over an unfamiliar repository and not understanding the AI's explanation
   The AI's response may contain many internal terms that it has not defined. wait-what requires it to provide the context first and then restate the explanation, including the missing assumptions.

2. The AI uses incorrect terminology in a repository with multiple modules
   When a repository has multiple bounded contexts, the AI may use terminology from another module to explain the current module. wait-what uses CONTEXT-MAP.md to find the CONTEXT.md file for the current module, and then restates the explanation using its standard terminology.

3. Long English technical explanations are difficult to read
   AI output may contain long sentences, many subordinate clauses, and frequent passive voice. wait-what requires the AI to use ASD-STE100 Simplified Technical English instead: one idea per sentence, consistent terminology, and short sentences.

Core Functions

1. Add context and explain again
   It does not repeat the previous response word for word. Instead, it first explains the background and assumptions, and then reorganizes the explanation. This is useful when the previous response skipped steps or omitted too much information.

2. Switch to ASD-STE100 Simplified Technical English
   ASD-STE100 is a controlled-language standard for technical documentation. It restricts vocabulary and sentence structures. The skill requires the AI to rewrite the explanation according to these rules so that non-native English speakers can understand it on the first read.

3. Use the project's standard terminology
   It prioritizes the project terms defined in CONTEXT.md instead of the model's usual wording. When the repository has multiple context files, it first reads CONTEXT-MAP.md to locate the correct one.

Frequently Asked Questions

What does the wait-what skill do?

It does one thing: it explains again the AI response that you found unclear. During the restatement, it adds the missing context, uses Simplified Technical English, and prioritizes the terminology defined in your project's CONTEXT.md. It does not answer new questions or modify code.

What is ASD-STE100 Simplified Technical English?

ASD-STE100 is a controlled-language standard for technical documentation. It uses a limited vocabulary and fixed sentence structures. Each sentence expresses only one idea to reduce ambiguity. This skill makes the AI rewrite its response according to these rules instead of using unrestricted language.

What is the relationship between CONTEXT.md and CONTEXT-MAP.md?

CONTEXT.md stores the standard terminology for a specific context, such as a module or domain. It defines the precise meaning of each term in that context. When a repository has multiple contexts, CONTEXT-MAP.md serves as an index. It tells you which context the current work belongs to and which CONTEXT.md file to read. When the repository has only one context, use that CONTEXT.md file directly.

Why can this skill not trigger automatically?

The skill is configured with disable-model-invocation: true. You must invoke it manually. The model cannot trigger it by itself. This design exists because only you know whether you understood the explanation. Automatic detection could cause unwanted triggers. For the same reason, the skill is not intended for repeated use in one conversation. Invoke it once, read the restatement, and then decide whether to use it again.

Will this skill modify my code or files?

No. It is a prompt-only skill. It only requires the AI to change how it explains something. It does not read or write files, execute commands, or modify code. It reads CONTEXT.md and CONTEXT-MAP.md only to use the correct terminology in the restatement. This also means that your project should ideally already contain these files before you use the skill. Otherwise, the restatement can only add context and simplify the English.

How is wait-what different from simply saying “Say that again”?

When you say “Say that again,” the AI often replaces a few words with synonyms and makes the original explanation even harder to understand. wait-what defines three fixed requirements: add context, use Simplified Technical English, and use the standard terminology from CONTEXT.md. The clear constraints make the restatement more consistent.