moai-adk 0.9.0__py3-none-any.whl → 0.15.0__py3-none-any.whl

moai_adk/templates/.claude/commands/alfred/0-project.md

@@ -1,10 +1,6 @@
 ---
 name: alfred:0-project
-description: "Initialize project document - create product/structure/tech.md and set optimization for each language"
-# Translations:
-# - ko: "프로젝트 문서 초기화 - product/structure/tech.md 생성 및 언어별 최적화 설정"
-# - ja: "プロジェクト文書の初期化 - product/structure/tech.mdの作成と言語別最適化設定"
-# - zh: "初始化项目文档 - 创建product/structure/tech.md并设置语言优化"
+description: "Initialize project metadata and documentation"
 allowed-tools:
 - Read
 - Write
@@ -79,7 +75,7 @@ The user executes the `/alfred:8-project` command to start analyzing the project
 - ❌ Create a file in the `.claude/memory/` directory
 - ❌ Create a file `.claude/commands/alfred/*.json`
 - ❌ Unnecessary overwriting of existing documents
-- ❌ Date and numerical prediction (“within 3 months”, “50% reduction”) etc.)
+- ❌ Date and numerical prediction ("within 3 months", "50% reduction") etc.)
 - ❌ Hypothetical scenarios, expected market size, future technology trend predictions
 
 **Expressions to use**:
@@ -96,6 +92,8 @@ The user executes the `/alfred:8-project` command to start analyzing the project
 
 **목적**: 프로젝트 초기화 시작 전에 대화 언어를 설정하고 사용자 닉네임을 등록합니다. 이 설정은 모든 Alfred 프롬프트, 인터뷰 질문 및 생성된 문서에 적용됩니다.
 
+**UX 개선**: 2개 질문을 **1회 배치 호출**로 통합 (50% 상호작용 감소: 2 turns → 1 turn)
+
 ### 0.0 Alfred 자기소개 및 환영 인사
 
 Alfred가 첫 상호작용으로 다음과 같이 인사합니다:
@@ -108,79 +106,354 @@ MoAI-ADK의 SuperAgent로서 당신의 프로젝트를 함께 만들어갈 준
 먼저 기본 설정을 진행하겠습니다.
 ```
 
-### 0.1 언어 선택
+### 0.1 배치 설계: 언어 선택 + 사용자 닉네임 + GitHub 설정 확인 (1-3회 호출)
+
+Alfred가 `AskUserQuestion tool (documented in moai-alfred-interactive-questions skill)` 를 사용하여 **배치 호출**로 필수 정보를 수집합니다:
+
+**기본 배치 (항상 실행)**:
+- 언어 선택
+- 사용자 닉네임
 
-Alfred가 `AskUserQuestion tool (documented in moai-alfred-interactive-questions skill)` 를 사용하여 **첫 번째 상호작용**으로 언어 선택 메뉴를 표시합니다:
+**추가 배치 (팀 모드 감지 시)**:
+- GitHub "Automatically delete head branches" 설정 확인
 
-**Question**:
+#### 0.1.1 팀 모드 감지
+
+```bash
+# config.json에서 mode 확인
+grep "mode" .moai/config.json
+
+# 결과: "mode": "team" → 추가 질문 포함
+#      "mode": "personal" → 기본 질문만 실행
+```
+
+#### 0.1.2 기본 배치: 언어 선택 + 에이전트 프롬프트 언어 + 닉네임 (3개 질문, 1회 배치 호출)
+
+**배치 설계**: 3개 질문을 1회 호출로 통합 (UX 개선: 3 turns → 1 turn)
+
+**Example AskUserQuestion Call**:
+```python
+AskUserQuestion(
+    questions=[
+        {
+            "question": "Which language would you like to use for the project initialization and documentation?",
+            "header": "Language",
+            "multiSelect": false,
+            "options": [
+                {
+                    "label": "🌍 English",
+                    "description": "All dialogs and documentation in English"
+                },
+                {
+                    "label": "🇰🇷 한국어",
+                    "description": "All dialogs and documentation in Korean"
+                },
+                {
+                    "label": "🇯🇵 日本語",
+                    "description": "All dialogs and documentation in Japanese"
+                },
+                {
+                    "label": "🇨🇳 中文",
+                    "description": "All dialogs and documentation in Chinese"
+                }
+            ]
+        },
+        {
+            "question": "In which language should Alfred's sub-agent prompts be written?",
+            "header": "Agent Prompt Language",
+            "multiSelect": false,
+            "options": [
+                {
+                    "label": "🌐 English (Global Standard)",
+                    "description": "All sub-agent prompts in English for global consistency and team collaboration. Recommended for Claude Pro $20 users: reduces token usage by ~15-20%, lowering API costs"
+                },
+                {
+                    "label": "🗣️ Selected Language (Localized)",
+                    "description": "All sub-agent prompts in the language you selected above for local team efficiency"
+                }
+            ]
+        },
+        {
+            "question": "How would you like to be called in our conversations? (e.g., GOOS, Team Lead, Developer, or custom name - max 20 chars)",
+            "header": "Nickname",
+            "multiSelect": false,
+            "options": [
+                {
+                    "label": "Enter custom nickname",
+                    "description": "Type your preferred name using the 'Other' option below"
+                }
+            ]
+        }
+    ]
+)
 ```
-Which language would you like to use for the project initialization and documentation?
+
+**응답 처리**:
+
+**Q1 (사용자 언어)**:
+- Selected option stored as: `conversation_language: "ko"` (or "en", "ja", "zh", etc.)
+
+**Q2 (에이전트 프롬프트 언어)** - **NEW**:
+- **"English (Global Standard)"** → `agent_prompt_language: "english"`
+  - All sub-agent prompts written in English
+  - Recommended for global teams, code consistency, and international collaboration
+  - Project-manager, spec-builder, code-builder use English prompts internally
+- **"Selected Language (Localized)"** → `agent_prompt_language: "localized"`
+  - All sub-agent prompts written in the user-selected language
+  - Recommended for local teams, local documentation, and native language efficiency
+  - Project-manager receives prompts in selected language (e.g., Korean, Japanese)
+
+**Q3 (사용자 닉네임)**:
+- Custom nickname stored as: `user.nickname: "GOOS"` (or custom input)
+
+#### 0.1.3 팀 모드 추가 배치: GitHub 설정 & Git 워크플로우 선택 (팀 모드만)
+
+**조건**: `config.json`에서 `"mode": "team"` 감지 시 실행
+
+**배치 구성**: 2개 질문 (1회 호출로 통합)
+
+**Example AskUserQuestion Call**:
+```python
+AskUserQuestion(
+    questions=[
+        {
+            "question": "[Team Mode] Is 'Automatically delete head branches' enabled in your GitHub repository settings?",
+            "header": "GitHub Branch Settings",
+            "multiSelect": false,
+            "options": [
+                {
+                    "label": "✅ Yes, already enabled",
+                    "description": "PR merge 후 자동으로 원격 브랜치 삭제됨"
+                },
+                {
+                    "label": "❌ No, not enabled (Recommended: Enable)",
+                    "description": "Settings → General → '자동 삭제' 체크박스 확인 필요"
+                },
+                {
+                    "label": "🤔 Not sure / Need to check",
+                    "description": "GitHub Settings → General 확인 후 다시 진행"
+                }
+            ]
+        },
+        {
+            "question": "[Team Mode] Which Git workflow should we use when creating SPEC documents?",
+            "header": "SPEC Git Workflow",
+            "multiSelect": false,
+            "options": [
+                {
+                    "label": "📋 Feature Branch + PR",
+                    "description": "매 SPEC마다 feature 브랜치 생성 → PR 리뷰 → develop 병합. 팀 협업과 코드 리뷰에 최적"
+                },
+                {
+                    "label": "🔄 Direct Commit to Develop",
+                    "description": "브랜치 생성 없이 develop에 직접 커밋. 빠른 프로토타이핑과 단순 워크플로우에 최적"
+                },
+                {
+                    "label": "🤔 Decide per SPEC",
+                    "description": "SPEC 생성 시마다 매번 선택. 유연성이 높지만 매번 결정 필요"
+                }
+            ]
+        }
+    ]
+)
 ```
 
-**Options** (AskUserQuestion with moai-alfred-interactive-questions):
-- **English** (en) — All dialogs and documentation in English
-- **한국어** (ko) — All dialogs and documentation in Korean
-- **日本語** (ja) — All dialogs and documentation in Japanese
-- **中文** (zh) — All dialogs and documentation in Chinese
-- **Other** — User can specify custom language (e.g., "Español", "Français", "Deutsch")
+**응답 처리**:
 
-### 0.2 Store Language Preference
+**Q1 (GitHub 설정)**:
+- **"Yes, already enabled"** → `auto_delete_branches: true` 저장
+- **"No, not enabled"** → `auto_delete_branches: false` + 권장사항 저장
+- **"Not sure"** → `auto_delete_branches: null` + 경고 메시지
 
-Alfred records the selected language:
+**Q2 (Git 워크플로우)**:
+- **"Feature Branch + PR"** → `spec_git_workflow: "feature_branch"` 저장
+  - `/alfred:1-plan` 실행 시 자동으로 feature 브랜치 생성
+  - git-manager가 PR 기반 워크플로우 적용
+- **"Direct Commit to Develop"** → `spec_git_workflow: "develop_direct"` 저장
+  - `/alfred:1-plan` 실행 시 develop 브랜치에 직접 커밋
+  - 브랜치 생성 과정 생략
+- **"Decide per SPEC"** → `spec_git_workflow: "per_spec"` 저장
+  - `/alfred:1-plan` 실행 시마다 git-manager가 사용자에게 선택 요청
 
+**User Response Example**:
+```
+Selected Language: 🇰🇷 한국어
+Selected Nickname: GOOS (typed via "Other" option)
+```
+
+---
+
+### 0.1.4 Domain Selection (Optional - All Modes)
+
+**Purpose**: Identify project domains to activate domain-expert agents for specialized guidance.
+
+**When to ask**: After language/nickname/GitHub settings complete
+
+**Batched Design**: Domain selection integrated into initial batch OR asked separately based on user preference
+
+**Example AskUserQuestion Call**:
+```python
+AskUserQuestion(
+    questions=[
+        {
+            "question": "Which domains does your project involve? (Select all that apply)",
+            "header": "Project Domains",
+            "multiSelect": true,
+            "options": [
+                {
+                    "label": "🎨 Frontend",
+                    "description": "React, Vue, Angular, Next.js, Nuxt, SvelteKit, Astro, Remix, SolidJS"
+                },
+                {
+                    "label": "⚙️ Backend",
+                    "description": "FastAPI, Flask, Django, Express, Fastify, NestJS, Spring Boot, Gin, Axum"
+                },
+                {
+                    "label": "🚀 DevOps",
+                    "description": "Railway, Vercel, Docker, Kubernetes, AWS, GCP, Azure, CI/CD"
+                },
+                {
+                    "label": "🗄️ Database",
+                    "description": "PostgreSQL, MySQL, MongoDB, Redis, database design and optimization"
+                },
+                {
+                    "label": "📊 Data Science",
+                    "description": "Data analysis, machine learning, data pipelines, notebooks"
+                },
+                {
+                    "label": "📱 Mobile",
+                    "description": "React Native, Flutter, iOS, Android app development"
+                },
+                {
+                    "label": "⚡ Skip",
+                    "description": "No domain selection (can add later via /alfred:1-plan)"
+                }
+            ]
+        }
+    ]
+)
+```
+
+**Response Processing**:
 ```json
 {
-  "conversation_language": "ko",
-  "conversation_language_name": "한국어",
-  "selected_at": "2025-10-22T12:34:56Z"
+  "stack": {
+    "selected_domains": ["frontend", "backend"],
+    "domain_selection_date": "2025-10-23T12:34:56Z"
+  }
 }
 ```
 
-This language preference is:
-- Passed to all sub-agents as a context parameter
-- Stored in `.moai/config.json` under `language` field
-- Used to generate all documentation in the selected language
-- Displayed in CLAUDE.md under "## Project Information"
+**Hint to User**: "You can always modify domains later during `/alfred:1-plan` when creating SPECs"
 
-### 0.2.5 사용자 닉네임 선택
+**Domain Expert Activation**:
+- Selected domains stored in `.moai/config.json`
+- Domain-expert agents activated during `/alfred:1-plan` (automatic keyword detection)
+- Domain-expert agents available as advisors during `/alfred:2-run`
+- Domain-specific sync routing enabled in `/alfred:3-sync`
 
-언어 선택 완료 후, Alfred가 `AskUserQuestion tool (documented in moai-alfred-interactive-questions skill)` 를 사용하여 사용자 닉네임을 요청합니다:
+---
 
-**질문**:
-```
-앞으로 대화에서 당신을 어떻게 부르면 좋을까요?
-(예: GOOS, 팀장님, 개발자님, 또는 자유롭게 입력)
+### 0.2 사용자 정보 저장
+
+Alfred가 선택된 언어, 닉네임, 그리고 팀 모드 설정을 다음과 같이 저장합니다:
+
+#### 0.2.1 기본 정보 저장 (항상) - 에이전트 프롬프트 언어 추가
+
+```json
+{
+  "language": {
+    "conversation_language": "ko",
+    "conversation_language_name": "한국어",
+    "agent_prompt_language": "localized",
+    "agent_prompt_language_description": "All sub-agent prompts written in the selected language (localized)"
+  },
+  "user": {
+    "nickname": "GOOS",
+    "selected_at": "2025-10-23T12:34:56Z"
+  },
+  "stack": {
+    "selected_domains": ["frontend", "backend"],
+    "domain_selection_date": "2025-10-23T12:34:56Z"
+  }
+}
 ```
 
-**입력 방식**:
-- 텍스트 직접 입력 가능 (자유 형식)
-- 예시: "GOOS", "팀장", "개발자" 등
-- 최대 20자 한도
+**에이전트 프롬프트 언어 옵션**:
+
+- **`"english"`** (Global Standard) - **💰 Claude Pro $20 사용자 추천**:
+  - All sub-agent prompts and internal communication in English
+  - Best for: International teams, global collaboration, code consistency
+  - Impact: Project-manager, spec-builder, code-builder all use English task prompts
+  - **Cost Benefit**: Reduces token usage by ~15-20% compared to non-English prompts
+    - English prompts are more efficient and use fewer tokens
+    - Significant cost savings for continuous API usage
+    - Example: 100,000 tokens in English ≈ 115,000-120,000 tokens in Korean/Japanese
+
+- **`"localized"`** (Localized - Default for non-English):
+  - All sub-agent prompts and internal communication in selected language
+  - Best for: Local teams, native language efficiency, culturally-specific guidance
+  - Impact: Project-manager, spec-builder, code-builder all use localized task prompts
+  - Note: Uses ~15-20% more tokens due to language characteristics
 
-### 0.2.6 사용자 정보 저장
+#### 0.2.2 GitHub & Git 워크플로우 설정 저장 (팀 모드만)
+
+**팀 모드 감지 시 추가 저장 - Feature Branch + PR 선택 시**:
+```json
+{
+  "github": {
+    "auto_delete_branches": true,
+    "spec_git_workflow": "feature_branch",
+    "checked_at": "2025-10-23T12:34:56Z",
+    "workflow_recommendation": "Feature branch를 사용한 PR 기반 협업 워크플로우. 매 SPEC마다 feature/spec-* 브랜치 생성, PR 리뷰 후 develop 병합"
+  }
+}
+```
 
-Alfred가 선택된 닉네임을 다음과 같이 저장합니다:
+**또는 - Direct Commit to Develop 선택 시**:
+```json
+{
+  "github": {
+    "auto_delete_branches": false,
+    "spec_git_workflow": "develop_direct",
+    "checked_at": "2025-10-23T12:34:56Z",
+    "workflow_recommendation": "develop 브랜치에 직접 커밋하는 단순 워크플로우. 브랜치 생성 과정 생략, 빠른 개발 속도"
+  }
+}
+```
 
+**또는 - Decide per SPEC 선택 시**:
 ```json
 {
-  "conversation_language": "ko",
-  "conversation_language_name": "한국어",
-  "user_nickname": "GOOS",
-  "selected_at": "2025-10-23T12:34:56Z"
+  "github": {
+    "auto_delete_branches": true,
+    "spec_git_workflow": "per_spec",
+    "checked_at": "2025-10-23T12:34:56Z",
+    "workflow_recommendation": "SPEC 생성 시마다 워크플로우 선택. /alfred:1-plan 실행 시 git-manager가 선택 요청"
+  }
 }
 ```
 
+#### 0.2.3 저장된 정보 활용
+
 이 정보는:
 - 모든 sub-agents 에게 컨텍스트 파라미터로 전달됨
-- `.moai/config.json` 의 `user` 필드에 저장됨
-- CLAUDE.md의 `{{USER_NICKNAME}}` 변수로 치환됨
+- `.moai/config.json` 의 `language`, `user`, `github` 필드에 저장됨
+- CLAUDE.md의 `{{CONVERSATION_LANGUAGE}}` 및 `{{USER_NICKNAME}}` 변수로 치환됨
 - 모든 Alfred 대화에서 사용됨
+- **팀 모드**: git-manager가 다음 워크플로우를 자동으로 적용:
+  - **`spec_git_workflow: "feature_branch"`**: `/alfred:1-plan` 실행 시 feature/spec-* 브랜치 생성, PR 기반 리뷰 프로세스 적용
+  - **`spec_git_workflow: "develop_direct"`**: `/alfred:1-plan` 실행 시 develop 브랜치에 직접 커밋, 브랜치 생성 과정 생략
+  - **`spec_git_workflow: "per_spec"`**: `/alfred:1-plan` 실행 시마다 사용자에게 워크플로우 선택 요청
 
-**예시**:
-```
-안녕하세요, GOOS님! 👋
+**설정 완료 출력 예시**:
+```markdown
+✅ 초기 설정 완료!
+
+언어: 한국어 (ko)
+닉네임: GOOS
 
-이제 프로젝트 환경 분석으로 진행하겠습니다...
+이제 GOOS님의 프로젝트 환경 분석으로 진행하겠습니다...
 ```
 
 ### 0.3 STEP 1로 전환
@@ -192,16 +465,6 @@ Alfred가 선택된 닉네임을 다음과 같이 저장합니다:
 - 생성된 문서 (product.md, structure.md, tech.md)가 선택된 언어로 작성됨
 - CLAUDE.md가 선택된 언어와 사용자 닉네임을 표시함
 
-**한국어 선택 시 출력 예시**:
-```markdown
-✅ 설정 완료!
-
-언어: 한국어 (ko)
-닉네임: GOOS
-
-이제 GOOS님의 프로젝트 환경 분석으로 진행하겠습니다...
-```
-
 ---
 
 ## 🚀 STEP 1: Environmental analysis and interview plan development
@@ -225,9 +488,9 @@ grep "optimized" .moai/config.json
 **Backup existence conditions**:
 - `.moai-backups/` directory exists
 - `.moai/project/*.md` file exists in the latest backup folder
-- `optimized: false` in `config.json` (immediately after reinitialization)
+- User's existing project files can be merged (regardless of optimized flag)
 
-**Select user if backup exists**  
+**Select user if backup exists**
 Call `AskUserQuestion tool (documented in moai-alfred-interactive-questions skill)` to display a TUI with the following options:
 - **Merge**: Merge backup contents and latest template (recommended)
 - **New**: Ignore the backup and start a new interview
@@ -238,12 +501,12 @@ Call `AskUserQuestion tool (documented in moai-alfred-interactive-questions skil
 - **"Create new"** → Proceed to Phase 1.2 (Project environment analysis) (existing process)
 - **"Skip"** → End task
 
-**No backup or optimized: true**:
+**No backup**:
 - Proceed directly to Phase 1.2 (project environment analysis)
 
 ---
 
-### 1.1 Backup merge workflow (when user selects “Merge”)
+### 1.1 Backup merge workflow (when user selects "Merge")
 
 **Purpose**: Restore only user customizations while maintaining the latest template structure.
 
@@ -391,7 +654,7 @@ Set optimization flags after the merge is complete:
  - Monolingual vs. hybrid vs. microservice
  - Code base size estimation
 
-### 1.3 Establish interview strategy (when user selects “New”)
+### 1.3 Establish interview strategy (when user selects "New")
 
 **Select question tree by project type**:
 
@@ -406,7 +669,7 @@ Set optimization flags after the merge is complete:
 - **Technical Questions**: Language/Framework, Quality Policy, Deployment Strategy
 - **Governance**: Security Requirements, Traceability Strategy (Optional)
 
-### 1.4 Generate Interview Plan Report (when user selects “Create New”)
+### 1.4 Generate Interview Plan Report (when user selects "Create New")
 
 **Format of plan to be presented to users**:
 
@@ -438,7 +701,7 @@ Set optimization flags after the merge is complete:
 
 ---
 **Approval Request**: Would you like to proceed with the interview using the above plan?
- (Choose “Proceed,” “Modify [Content],” or “Abort”)
+ (Choose "Proceed," "Modify [Content]," or "Abort")
 ```
 
 ### 1.5 Wait for user approval (moai-alfred-interactive-questions) (when user selects "New")
@@ -455,7 +718,7 @@ After Alfred receives the project-manager's interview plan report, uses `AskUser
 
 ---
 
-## 🚀 STEP 2: Execute project initialization (after user approves “New”)
+## 🚀 STEP 2: Execute project initialization (after user approves "New")
 
 **Note**: This step will only be executed if the user selects **"New"**.
 - When selecting "Merge": End the task in Phase 1.1 (Merge Backups)
@@ -471,49 +734,85 @@ Alfred starts project initialization by calling the project-manager agent with t
 **Parameters passed to project-manager**:
 - **conversation_language** (from STEP 0): Language code selected by user (e.g., "ko", "en", "ja", "zh")
 - **language_name** (from STEP 0): Display name of selected language (e.g., "Korean", "English")
+- **agent_prompt_language** (from STEP 0.1.2) - **NEW**:
+  - `"english"` = All sub-agent prompts in English (Global Standard)
+  - `"localized"` = All sub-agent prompts in selected conversation_language (Localized)
 - Detected Languages: [Language List from codebase detection]
 - Project Type: [New/Existing]
 - Existing Document Status: [Existence/Absence]
 - Approved Interview Plan: [Plan Summary]
+- **Team Mode Git Workflow** (from STEP 0.1.3):
+  - `spec_git_workflow: "feature_branch" | "develop_direct" | "per_spec"` (팀 모드만)
 
 **Execution**:
 ```
 Call the Task tool:
 - subagent_type: "project-manager"
 - description: "Initialize project with conversation language support"
-- prompt: """You are project-manager agent.
-
-LANGUAGE CONFIGURATION:
-- conversation_language: {{CONVERSATION_LANGUAGE}}
-- language_name: {{CONVERSATION_LANGUAGE_NAME}}
-
-PROJECT_TYPE: [new|existing]
-DETECTED_LANGUAGES: [detected codebase languages]
-
-CRITICAL INSTRUCTION:
-All interviews and generated documentation MUST be in conversation_language:
-- product.md: Generate in {{CONVERSATION_LANGUAGE}}
-- structure.md: Generate in {{CONVERSATION_LANGUAGE}}
-- tech.md: Generate in {{CONVERSATION_LANGUAGE}}
-
-If conversation_language is 'ko': All narrative content in Korean
-If conversation_language is 'ja': All narrative content in Japanese
-If conversation_language is other: Follow the specified language
-
-After project initialization, update .moai/config.json with nested language structure:
+- prompt: """당신은 project-manager 에이전트입니다.
+
+언어 설정:
+- 대화_언어: {{CONVERSATION_LANGUAGE}} (모든 대화, 문서에 사용)
+- 언어명: {{CONVERSATION_LANGUAGE_NAME}}
+- 에이전트_프롬프트_언어: {{AGENT_PROMPT_LANGUAGE}} (내부 sub-agent 통신 언어)
+
+에이전트 프롬프트 언어에 따른 작업 방식:
+
+1. **agent_prompt_language = "english"** (Global Standard):
+   - 당신(project-manager)은 **영어**로 사고하고 작업합니다
+   - 모든 내부 분석과 계획을 영어로 진행합니다
+   - 생성된 product.md, structure.md, tech.md는 **{{CONVERSATION_LANGUAGE}}**로 작성합니다
+   - Sub-agent들(spec-builder 등)에게 전달하는 프롬프트는 **영어**입니다
+
+2. **agent_prompt_language = "localized"** (Localized):
+   - 당신(project-manager)은 **{{CONVERSATION_LANGUAGE}}**로 사고하고 작업합니다
+   - 모든 내부 분석과 계획을 {{CONVERSATION_LANGUAGE}}로 진행합니다
+   - 생성된 product.md, structure.md, tech.md는 **{{CONVERSATION_LANGUAGE}}**로 작성합니다
+   - Sub-agent들(spec-builder 등)에게 전달하는 프롬프트도 **{{CONVERSATION_LANGUAGE}}**입니다
+
+중요: 대화_언어(conversation_language)와 에이전트_프롬프트_언어(agent_prompt_language)는 다를 수 있습니다!
+- 대화_언어는 **사용자와의 대화**, **생성 문서**에 사용
+- 에이전트_프롬프트_언어는 **sub-agents 통신**, **내부 prompt**에 사용
+
+GIT 워크플로우 설정 (팀 모드):
+- spec_git_workflow: [feature_branch | develop_direct | per_spec]
+  - "feature_branch": feature/spec-* 브랜치 생성, PR 기반 리뷰, develop 병합
+  - "develop_direct": develop에 직접 커밋, 브랜치 생성 안 함
+  - "per_spec": SPEC별로 사용자에게 물어봄 (/alfred:1-plan 실행 중)
+- 참고: 이 값을 .moai/config.json github.spec_git_workflow에 저장하여 git-manager가 참조하도록
+
+프로젝트_타입: [new|existing]
+감지된_언어들: [감지된 코드베이스 언어들]
+
+중요 지시사항:
+모든 인터뷰와 생성된 문서는 대화_언어(conversation_language)로 작성되어야 합니다:
+- product.md: {{CONVERSATION_LANGUAGE}}로 생성
+- structure.md: {{CONVERSATION_LANGUAGE}}로 생성
+- tech.md: {{CONVERSATION_LANGUAGE}}로 생성
+
+conversation_language가 'ko'인 경우: 모든 설명 내용을 한국어로
+conversation_language가 'ja'인 경우: 모든 설명 내용을 일본어로
+다른 언어인 경우: 지정된 언어를 따릅니다
+
+프로젝트 초기화 후, 다음과 같이 .moai/config.json 업데이트:
 {
   "language": {
     "conversation_language": "{{CONVERSATION_LANGUAGE}}",
-    "conversation_language_name": "{{CONVERSATION_LANGUAGE_NAME}}"
+    "conversation_language_name": "{{CONVERSATION_LANGUAGE_NAME}}",
+    "agent_prompt_language": "{{AGENT_PROMPT_LANGUAGE}}"
+  },
+  "github": {
+    "spec_git_workflow": "[feature_branch|develop_direct|per_spec]"
   }
 }
 
-SKILL INVOCATION:
-Use explicit Skill() calls when needed:
-- Skill("moai-alfred-language-detection") for codebase language detection
-- Skill("moai-foundation-langs") for multi-language project setup
+스킬 호출:
+필요 시 명시적 Skill() 호출 사용:
+- Skill("moai-alfred-language-detection") - 코드베이스 언어 감지
+- Skill("moai-foundation-langs") - 다국어 프로젝트 설정
 
-TASK: Conduct project interviews and create/update product/structure/tech.md documents."""
+작업: 프로젝트 인터뷰를 진행하고 product/structure/tech.md 문서를 생성/업데이트합니다.
+에이전트_프롬프트_언어 설정에 따라 sub-agent들과의 통신 언어를 결정합니다."""
 ```
 
 **Outcome**: The project-manager agent conducts structured interviews entirely in the selected language and creates/updates product/structure/tech.md documents in that language.
@@ -526,7 +825,7 @@ After the project-manager has finished creating the document, **Alfred can optio
 
 | Conditions                           | Automatic selection Skill    | Purpose                                |
 | ------------------------------------ | ---------------------------- | -------------------------------------- |
-| User Requests “Quality Verification” | moai-alfred-trust-validation | Initial project structure verification |
+| User Requests "Quality Verification" | moai-alfred-trust-validation | Initial project structure verification |
 
 **Execution flow** (optional):
 ```
@@ -562,24 +861,24 @@ After the project-manager has finished creating the document, **Alfred can optio
 **Interview Flow**:
 
 1. **Product Discovery** (create product.md)
- - Define core mission (@DOC:MISSION-001)
- - Identify key user base (@SPEC:USER-001)
- - Identify key problems to solve (@SPEC:PROBLEM-001)
- - Summary of differences and strengths (@DOC:STRATEGY-001)
- - Setting success indicators (@SPEC:SUCCESS-001)
+ - Define core mission (DOC:MISSION-001)
+ - Identify key user base (SPEC:USER-001)
+ - Identify key problems to solve (SPEC:PROBLEM-001)
+ - Summary of differences and strengths (DOC:STRATEGY-001)
+ - Setting success indicators (SPEC:SUCCESS-001)
 
 2. **Structure Blueprint** (create structure.md)
- - Selection of architecture strategy (@DOC:ARCHITECTURE-001)
- - Division of responsibilities by module (@DOC:MODULES-001)
- - External system integration plan (@DOC:INTEGRATION-001)
- - Define traceability strategy (@DOC:TRACEABILITY-001)
+ - Selection of architecture strategy (DOC:ARCHITECTURE-001)
+ - Division of responsibilities by module (DOC:MODULES-001)
+ - External system integration plan (DOC:INTEGRATION-001)
+ - Define traceability strategy (DOC:TRACEABILITY-001)
 
 3. **Tech Stack Mapping** (written by tech.md)
- - Select language & runtime (@DOC:STACK-001)
- - Determine core framework (@DOC:FRAMEWORK-001)
- - Set quality gate (@DOC:QUALITY-001)
-   - Define security policy (@DOC:SECURITY-001)
- - Plan distribution channels (@DOC:DEPLOY-001)
+ - Select language & runtime (DOC:STACK-001)
+ - Determine core framework (DOC:FRAMEWORK-001)
+ - Set quality gate (DOC:QUALITY-001)
+   - Define security policy (DOC:SECURITY-001)
+ - Plan distribution channels (DOC:DEPLOY-001)
 
 **Automatically generate config.json**:
 ```json
@@ -686,7 +985,7 @@ Based on the collected information, it is reflected in three major documents:
  - Project mission extracted from existing README/document
  - Main user base and scenario inferred from code
  - Backtracking of core problem to be solved
- - Preservation of existing assets in “Legacy Context”
+ - Preservation of existing assets in "Legacy Context"
 
 2. Contents reflected in **structure.md**
  - Identified actual directory structure
@@ -703,7 +1002,7 @@ Based on the collected information, it is reflected in three major documents:
 
 **Preservation Policy**:
 - Supplement only the missing parts without overwriting existing documents
-- Preserve conflicting content in the “Legacy Context” section
+- Preserve conflicting content in the "Legacy Context" section
 - Mark items needing improvement with @CODE and TODO tags
 
 **Example Final Report**:
@@ -802,7 +1101,7 @@ Alfred only calls the trust-checker agent to perform project initial structural
 ❌ **Critical**: Needs fix
 - Required section missing
 - config.json syntax error
-- User choice: “Revalidate after fix” or “Skip”
+- User choice: "Revalidate after fix" or "Skip"
 
 **Skip verification**:
 - Verification is not run by default
@@ -829,11 +1128,11 @@ cc-manager selects the required sub-agents and skills based on the briefing.The
 
 | Project requirements (document basis)                                              | Recommended sub-agent/skill                                                                                             | Purpose                                                                |
 | ---------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------- |
-| High quality and coverage goals (`product.md@SPEC:SUCCESS-001`)                    | `tdd-implementer`, `moai-essentials-debug`, `moai-essentials-review`                                                    | Establishment of RED·GREEN·REFACTOR workflow                           |
-| Traceability/TAG improvement request (`structure.md@DOC:TRACEABILITY-001`)         | `doc-syncer`, `moai-alfred-tag-scanning`, `moai-alfred-trust-validation`                                                | Enhanced TAG traceability and document/code synchronization            |
+| High quality and coverage goals (product.md:SPEC:SUCCESS-001)                    | `tdd-implementer`, `moai-essentials-debug`, `moai-essentials-review`                                                    | Establishment of RED·GREEN·REFACTOR workflow                           |
+| Traceability/TAG improvement request (structure.md:DOC:TRACEABILITY-001)         | `doc-syncer`, `moai-alfred-tag-scanning`, `moai-alfred-trust-validation`                                                | Enhanced TAG traceability and document/code synchronization            |
 | Deployment automation/branch strategy required (`structure.md` Architecture/TODO)  | `git-manager`, `moai-alfred-git-workflow`, `moai-foundation-git`                                                        | Branch Strategy·Commit Policy·PR Automation                            |
 | Refactoring legacy modules (`product.md` BACKLOG, `tech.md` TODO)                  | `implementation-planner`, `moai-essentials-refactor`                                                                     | Technical Debt Diagnosis and Refactoring Roadmap                       |
-| Strengthening regulatory/security compliance (`tech.md@DOC:SECURITY-001`)          | `quality-gate`, `moai-alfred-trust-validation`, `moai-foundation-trust`, `moai-domain-security`                         | TRUST S (Secured) and Trackable Compliance, Security Consulting        |
+| Strengthening regulatory/security compliance (tech.md:DOC:SECURITY-001)          | `quality-gate`, `moai-alfred-trust-validation`, `moai-foundation-trust`, `moai-domain-security`                         | TRUST S (Secured) and Trackable Compliance, Security Consulting        |
 | CLI Automation/Tooling Requirements (`tech.md` BUILD/CLI section)                  | `implementation-planner`, `moai-domain-cli-tool`, detected language skills (e.g. `moai-lang-python`)                    | CLI command design, input/output standardization                       |
 | Data analysis/reporting needs (`product.md` DATA, `tech.md` ANALYTICS)             | `implementation-planner`, `moai-domain-data-science`, detected language skills                                          | Data Pipeline·Notebook Job Definition                                  |
 | Improved database structure (`structure.md` DB, `tech.md` STORAGE)                 | `doc-syncer`, `moai-domain-database`, `moai-alfred-tag-scanning`                                                        | Strengthening schema documentation and TAG-DB mapping                  |
@@ -852,9 +1151,9 @@ If multiple conditions are met, the candidates are merged without duplicates and
 
 #### 2.6.2 User confirmation flow
 
-`AskUserQuestion tool (documented in moai-alfred-interactive-questions skill)` asks “whether to enable recommended items.”
+`AskUserQuestion tool (documented in moai-alfred-interactive-questions skill)` asks "whether to enable recommended items."
 - Provides three options: **Install all** / **Install selectively** / **Do not install**.
-Selecting “Selective Install” presents the list of candidates again as multiple choices, allowing the user to select only the items they need.
+Selecting "Selective Install" presents the list of candidates again as multiple choices, allowing the user to select only the items they need.
 
 #### 2.6.3 Activation and Recording Steps
 
@@ -863,9 +1162,9 @@ Selecting “Selective Install” presents the list of candidates again as multi
 - Call `subagent_type: "cc-manager"` with the `Task` tool and include a briefing and user selections in the prompt.
 - cc-manager determines the necessary sub-agents and skills based on the briefing, and copies and updates `CLAUDE.md`, `.claude/agents/alfred/*.md`, and `.claude/skills/*.md` as customized for the project.
 3. **Check for configuration updates**: Review the results reflected by cc-manager.
-- Sub-Agents: Keep the `.claude/agents/alfred/` template active and list it in the `CLAUDE.md` “Agents” section.
-- Skills: Check the `.claude/skills/` document and add it to the `CLAUDE.md` “Skills” section.
-- Output style: Apply `.claude/output-styles/alfred/` and record the activation in `CLAUDE.md` “Output Styles”.
+- Sub-Agents: Keep the `.claude/agents/alfred/` template active and list it in the `CLAUDE.md` "Agents" section.
+- Skills: Check the `.claude/skills/` document and add it to the `CLAUDE.md` "Skills" section.
+- Output style: Apply `.claude/output-styles/alfred/` and record the activation in `CLAUDE.md` "Output Styles".
 4. **Update config.json**
    ```json
    {
@@ -878,17 +1177,17 @@ Selecting “Selective Install” presents the list of candidates again as multi
    }
    ```
 Merge existing properties, if any.
-5. **Final Report**: Add a list of “Activated Sub-Agents/Skills/Style” and a `cc_manager_briefing` summary at the top of the Completion Report, and reflect the same contents in the `CLAUDE.md` table so that they are automatically searched in subsequent commands.
+5. **Final Report**: Add a list of "Activated Sub-Agents/Skills/Style" and a `cc_manager_briefing` summary at the top of the Completion Report, and reflect the same contents in the `CLAUDE.md` table so that they are automatically searched in subsequent commands.
 
 ## Interview guide by project type
 
 ### New project interview area
 
 **Product Discovery** (product.md)
-- Core mission and value proposition 
- - Key user bases and needs 
- - 3 key problems to solve 
- - Differentiation compared to competing solutions 
+- Core mission and value proposition
+ - Key user bases and needs
+ - 3 key problems to solve
+ - Differentiation compared to competing solutions
  - Measurable indicators of success
 
 **Structure Blueprint** (structure.md)
@@ -936,14 +1235,14 @@ Merge existing properties, if any.
 
 **Error 1**: Project language detection failed
 ```
-Symptom: “Language not detected” message
+Symptom: "Language not detected" message
 Solution: Specify language manually or create language-specific settings file
 ```
 
 **Error 2**: Conflict with existing document
 ```
 Symptom: product.md already exists and has different contents
-Solution: Preserve existing contents and add new contents in “Legacy Context” section
+Solution: Preserve existing contents and add new contents in "Legacy Context" section
 ```
 
 **Error 3**: Failed to create config.json
@@ -996,8 +1295,8 @@ This subcommand is executed under the following conditions:
  - Set optimized=true
    ```
 
-4. **Waiting for user approval**  
-`AskUserQuestion tool (documented in moai-alfred-interactive-questions skill)` asks “Do you want to proceed with template optimization?” and provides the following options.
+4. **Waiting for user approval**
+`AskUserQuestion tool (documented in moai-alfred-interactive-questions skill)` asks "Do you want to proceed with template optimization?" and provides the following options.
 - **Proceed** → Phase 2 execution
 - **Preview** → Display change details and recheck
 - **Skip** → keep optimized=false
@@ -1160,8 +1459,8 @@ Alfred: Skill("moai-alfred-template-generator")
 **Users can skip Phase 3**:
 
 **Skip condition**:
-- User explicitly selects “Skip”
-- “Simple project” when Alfred automatically determines (only basic features required)
+- User explicitly selects "Skip"
+- "Simple project" when Alfred automatically determines (only basic features required)
 
 **Skip effect**:
 - Maintain all 37 skills (no lightweighting)
@@ -1181,6 +1480,43 @@ After initialization is complete:
 - **Set Change**: Run `/alfred:0-project` again to update document
 - **Template optimization**: Run `/alfred:0-project update` after `moai-adk update`
 
+## Final Step
+
+After project initialization completes, Alfred automatically invokes AskUserQuestion to ask the user what to do next:
+
+```python
+AskUserQuestion(
+    questions=[
+        {
+            "question": "프로젝트 초기화가 완료되었습니다. 다음으로 뭘 하시겠습니까?",
+            "header": "다음 단계",
+            "multiSelect": false,
+            "options": [
+                {
+                    "label": "📋 스펙 작성 진행",
+                    "description": "/alfred:1-plan 실행하여 첫 SPEC 작성"
+                },
+                {
+                    "label": "🔍 프로젝트 구조 검토",
+                    "description": "생성된 문서 검토 및 수정"
+                },
+                {
+                    "label": "🔄 새 세션 시작",
+                    "description": "성능 최적화를 위해 /clear 실행"
+                }
+            ]
+        }
+    ]
+)
+```
+
+**User Responses**:
+- **📋 스펙 작성 진행**: Proceed to `/alfred:1-plan` for creating first SPEC
+- **🔍 프로젝트 구조 검토**: Review and modify generated project documents
+- **🔄 새 세션 시작**: Execute `/clear` to start fresh session (recommended for performance)
+
+---
+
 ## Related commands
 
 - `/alfred:1-plan` - Start writing SPEC