Use only when the user explicitly requests it with /tk-mwhat or phrases such as 뭐라고?, 뭐라는 거야?, or 무슨 말이죠?. The input is only the immediately preceding explanation/source in the conversation. If none exists or the reference is ambiguous, output exactly Unverifiable and stop.
Output
Output only these two blocks.
🤔 쉽게 말하면
[뜻과 현재 문제를 최대 3개의 짧은 줄로 재설명]
💡 추천
[추천 또는 그대로 보낼 다음 문장 최대 2개의 짧은 줄]
Write more briefly than the source without changing its meaning. Preserve paths, commands, URLs, literals,
verified UI strings, Status/IDs, and original authorship attribution exactly. Keep a verified UI string's language,
case, punctuation, and spacing; do not translate, paraphrase, or normalize it. Do not turn an enum, code identifier,
or i18n key into a UI label. If there is no basis for a recommendation, use
추천: 없음.
After the re-explanation, hard stop. Do not modify files, implement anything, create plans or summaries, or run commands or tools. New tasks, change requests, and general summary requests are outside this skill; stop with NotApplicable.
Source
Prior precedent is TigerKit mwhat (commit c6963e8, skills/mwhat/SKILL.md):
two short Korean explanation blocks, exact source/literal preservation, and no implementation contract.
Matt Pocock's original wait-what was not available in this checkout or its records, so no behavior was inferred.