multi-agent-starter CLAUDE.md

MultiAgent Orchestration — Operating Rules

Architecture

Orchestrator (Claude Code session, internal reasoning)
└── Worker Pool (모두 외부 호출 — 승인 필요)
    ├── claude-main    [strategist] 기획 · 설계 · 아키텍처 · 전략 · 디자인 방향 · 문체 글쓰기 · 디버깅 원인 분석
    ├── codex-main     [engineer·computer-use] 대규모 구현 · 코드 분석 · 테스트 · diff · 로컬 검증 · 브라우저 자동화 · 이미지 생성
    ├── codex-critic   [reviewer] 산출물 리뷰·비평 (Codex의 주된 역할)
    └── gemini         [multimodal] 멀티모달 · 긴 문서 · 제3자 시각의 검토

능력 슬롯 → 워커 배정의 정본은 _shared/capability-profile.md(가변층 — 신모델 출시 시 프로필만 갱신).

중요: Orchestrator의 내부 추론은 worker가 아님. claude-main worker 호출은 별도 모델 호출이므로 승인·쿼터 대상.

운영 원칙 (Operating Principles)

Behavioral guidelines to reduce common LLM coding mistakes. Merge with project-specific instructions as needed.

Tradeoff: These guidelines bias toward caution over speed. For trivial tasks, use judgment.

1. Think Before Coding

Don't assume. Don't hide confusion. Surface tradeoffs.

Before implementing:

• State your assumptions explicitly. If uncertain, ask.
• If multiple interpretations exist, present them - don't pick silently.
• If a simpler approach exists, say so. Push back when warranted.
• If something is unclear, stop. Name what's confusing. Ask.

2. Simplicity First

Minimum code that solves the problem. Nothing speculative.

• No features beyond what was asked.
• No abstractions for single-use code.
• No "flexibility" or "configurability" that wasn't requested.
• No error handling for impossible scenarios.
• If you write 200 lines and it could be 50, rewrite it.

Ask yourself: "Would a senior engineer say this is overcomplicated?" If yes, simplify.

3. Surgical Changes

Touch only what you must. Clean up only your own mess.

When editing existing code:

• Don't "improve" adjacent code, comments, or formatting.
• Don't refactor things that aren't broken.
• Match existing style, even if you'd do it differently.
• If you notice unrelated dead code, mention it - don't delete it.