코드를 읽으면 알 수 있는 사실은 절대 CLAUDE.md에 쓰지 마십시오.
클로드는 코드를 읽을 수 있으므로, 코드로 드러나지 않는 함정과 규칙만 CLAUDE.md에 남깁니다.
Never write in CLAUDE.md anything Claude can infer from reading the code itself. Claude reads code — write only the traps and rules that the code can't reveal.
코드를 읽으면 알 수 있는 사실은 절대 CLAUDE.md에 쓰지 마십시오.
클로드는 코드를 읽을 수 있으므로, 코드로 드러나지 않는 함정과 규칙만 CLAUDE.md에 남깁니다.
좋은 CLAUDE.md와 나쁜 CLAUDE.md를 가르는 기준입니다. 나쁜 예: "React를 사용한다", "src 폴더에 컴포넌트가 있다", "TypeScript로 작성되었다" — 클로드가 코드를 직접 읽으면 알 수 있는 정보입니다. 좋은 예: 코드 어디에도 적혀 있지 않은 도메인 함정, 팀 컨벤션, 재현되는 실수 패턴 — 코드에서 추출 불가능한 정보만 가드레일로 남깁니다.Also in EnglishNever write in CLAUDE.md anything Claude can infer from reading the code itself. Claude reads code — write only the traps and rules that the code can't reveal.
The boundary that separates a good CLAUDE.md from a bad one. Bad: "we use React," "components live in src/," "written in TypeScript" — Claude derives all of this by reading the code. Good: domain traps, team conventions, recurring mistake patterns — guardrails for what the code can't tell Claude on its own.