hiro-inagawa/write-like-me

** 10 초에 그것을 시도. ** Repo를 복제하고 Markdown 또는 텍스트 파일에 체크러를 실행합니다. 프로필 없음, 설치 없음, Python 3.8+만:

python 스크립트/voice check.py your-draft.md

그것은 모든 em dash, 필러 구문, 헤지, AI는 라인 번호와 수정을 말한다, 그 때 verdict. 자신의 음성 프로파일을 구축하면, 당신이 그것을 원하면.

A Claude Code Skills that analyzes how you write and generates for Writing and revising prose. 그것은 언어 및 심리 측정 방법을 사용하여 주제 설명보다는 자신의 쓰기의 corpus에서 실제 스타일을 캡처합니다.

** v2의 새로운 (9 월 2026).** 기술이 이제 결정적인 검수원을 발송합니다. 초안이 작성된 후, python scripts/voice_check.py draft.md --profile voices/ /profile.json]는 라인 번호와 수정을 가진 모든 위반을 보고, 단단한 규칙에 비소를 종료합니다. 따라서 음성은 모델이 주의해야 할 모델이 아닌 코드에 의해 시행됩니다. 각 음성은 또한 기계 읽기 쉬운 profile.json와 황금 시험 세트를 얻고, eval 마구 문은 각 규칙 변화합니다. 자세한 내용은 섹션에서 "확인 초안"및 "확인"아래에서, 및 CHANGELOG.md.

Try it in ten seconds. Clone the repo and run the checker on any Markdown or text file. No profile, no installs, Python 3.8+ only:

python scripts/voice_check.py your-draft.md

It prints every em dash, filler phrase, hedge, and AI tell with a line number and a fix, then a verdict. Building your own voice profile comes after, if you want it.

A Claude Code skill that analyzes how you write and generates a personalized voice profile for writing and revising prose. It uses linguistic and psychological measurement methods rather than subjective descriptions to capture your actual style from a corpus of your own writing.

New in v3.2 (November 2026). German. A voice can now be German (Hochdeutsch). The checker knows the AI habits of German text: the English tells in German words, and the ones that only exist in German, such as English quotation marks, English number formats, word-for-word translations from English and a text that mixes du and Sie. The measurements use German function words, German readability formulas and the German spaCy model. Details are in the section “German” below. In German the skill can be called “Schreib wie ich”.

New in v3 (October 2026). The skill now teaches the method as well as measuring the result. For every text that a reader will see, it works in two messages. The first explains the subject to the author in plain paragraphs. The second, sent after approval, is the text, cut from that explanation. The checker gained six reader-text rules, a check that the text contains no number or name that the explanation lacked, and a check that a rejected phrase does not return. Claude Code hooks run the method and the checker without a reminder. Details are in the sections “The two messages”, “New checks in v3”, “Hooks for Claude Code” and “Provisional rules”, and in CHANGELOG.md.

New in v2 (September 2026). The skill now ships a deterministic checker. After a draft is written, python scripts/voice_check.py draft.md --profile voices//profile.json reports every violation with a line number and a fix, and exits non-zero on a hard rule, so the voice is enforced by code instead of by asking the model to be careful. Each voice also gets a machine-readable profile.json and a golden test set, and an eval harness gates every rule change. Details are in the sections “Checking a draft” and “Evaluating the checker” below, and in CHANGELOG.md.

The same skill supports multiple voices. You can build one from your primary writing register, another from a different register, and another from an author whose style you want to study, and switch between them within the same skill.

---