@yeongjaeyou/claude-code-config 0.18.5 → 0.18.7

package/.claude/commands/gh/auto-review-loop.md

@@ -51,7 +51,7 @@ MAX_ITERATIONS=10
 
 ### Step 2: Wait for All Checks
 
-`gh pr checks`로 CodeRabbit과 Actions 상태를 한 번에 확인.
+Check CodeRabbit and Actions status at once with `gh pr checks`.
 
 ```bash
 wait_for_all_checks() {
@@ -81,7 +81,7 @@ wait_for_all_checks() {
 
 ### Step 3: Check Unresolved Threads (GraphQL)
 
-CodeRabbit 리뷰 코멘트 중 미해결 건수 확인. This is the ground truth.
+Check unresolved CodeRabbit review comment count. This is the ground truth.
 
 ```bash
 check_unresolved_threads() {

package/.claude/commands/gh/decompose-issue.md

@@ -14,9 +14,9 @@ Break down large work items into manageable, independent issues. Follow project
    - Detect test frameworks (jest.config.*, pytest.ini, vitest.config.*, pyproject.toml, etc.)
 3. **Check TDD applicability** (if user hasn't specified):
    - Analyze work type: code implementation vs docs/infra/config
-   - If test framework detected + code work → Ask: "TDD 기반으로 이슈 생성하시겠어요?"
-   - If no test framework → Inform: "TDD 반영이 필요없습니다. (이유: 테스트 프레임워크 미감지)"
-   - If non-code work (docs/infra) → Inform: "TDD 반영이 필요없습니다. (이유: 비코드 작업)"
+   - If test framework detected + code work → Ask: "Create issues with TDD approach?"
+   - If no test framework → Inform: "TDD not required. (Reason: No test framework detected)"
+   - If non-code work (docs/infra) → Inform: "TDD not required. (Reason: Non-code work)"
    - If TDD selected: Add `<!-- TDD: enabled -->` marker to each issue body
 4. Analyze work: Understand core requirements and objectives
 5. Decompose work: Split major tasks into smaller, manageable sub-tasks or issues. **Aim for optimal count over excessive issues (keep it manageable)**
@@ -164,14 +164,14 @@ Always include an ASCII diagram in the issue body to visualize the execution flo
 
 ## Verification Guidelines
 
-이슈 작업 완료 시 반드시 실행 검증 수행:
+Verification is mandatory when issue work is complete:
 
-| 작업 유형 | 검증 방법 |
-|-----------|-----------|
-| Python 코드 | `python -m py_compile file.py` + 실제 실행 |
-| TypeScript/JS | `tsc --noEmit` 또는 빌드 |
-| API/서버 | 엔드포인트 호출 테스트 |
-| CLI 도구 | 기본 명령어 실행 |
-| 설정 파일 | 관련 도구로 로드 확인 |
+| Work Type | Verification Method |
+|-----------|---------------------|
+| Python code | `python -m py_compile file.py` + actual execution |
+| TypeScript/JS | `tsc --noEmit` or build |
+| API/Server | Endpoint call test |
+| CLI tools | Run basic commands |
+| Config files | Verify loading with related tools |
 
-**파일만 생성하고 실행 검증 없이 완료 처리 금지**
+**Never mark complete if only files are created without execution verification**

package/.claude/commands/gh/resolve-and-review.md

@@ -4,7 +4,7 @@ description: Full issue resolution pipeline with automated code review
 
 # Resolve and Review
 
-Issue 해결부터 CodeRabbit 리뷰 처리까지 자동화된 전체 파이프라인.
+Automated full pipeline from issue resolution to CodeRabbit review handling.
 
 ## Arguments
 

package/.claude/guidelines/id-reference.md

@@ -20,7 +20,7 @@ Reference for different ID formats used across GitHub and TaskMaster.
 
 ## Linking GitHub and TaskMaster
 
-- PR body: `Task Master 참조: task N` or `TaskMaster: task N`
+- PR body: `TaskMaster: task N` or `Task Master ref: task N`
 - Issue closing: `Closes #N` (GitHub Issue)
 
 ## Example
@@ -28,7 +28,7 @@ Reference for different ID formats used across GitHub and TaskMaster.
 ```
 PR #13: "feat: Add authentication"
 ├── Closes #1 (GitHub Issue)
-└── Task Master 참조: task 1
+└── TaskMaster: task 1
     ├── task 1.1 (subtask)
     └── task 1.2 (subtask)
 ```

package/.claude/guidelines/prd-guide.md

@@ -1,36 +1,36 @@
-# PRD 초안 템플릿
+# PRD Draft Template
 
-여기에 아이디어를 자유롭게 작성하세요.
-`/tm:convert-prd` 커맨드로 TaskMaster PRD 포맷으로 변환됩니다.
+Write your ideas freely here.
+Use `/tm:convert-prd` command to convert to TaskMaster PRD format.
 
 ---
 
-## 뭘 만들고 싶은지
+## What to Build
 
-(프로젝트 목적, 해결하려는 문제)
+(Project purpose, problem to solve)
 
 
 
-## 주요 기능
+## Key Features
 
-(만들고 싶은 기능들 나열)
+(List of features to build)
 
 
 
-## 기술 스택
+## Tech Stack
 
-(사용할 기술, 프레임워크 등)
+(Technologies, frameworks to use)
 
 
 
-## 만드는 순서
+## Build Order
 
-(먼저 만들 것, 나중에 만들 것)
+(What to build first, what to build later)
 
 
 
-## 기타
+## Notes
 
-(걱정되는 점, 참고사항 등)
+(Concerns, references, etc.)
 
 

package/.claude/guidelines/work-guidelines.md

@@ -23,7 +23,7 @@ Write like a human, not a chatbot. Applies to ALL text: responses, documentation
 - Over-explanation: Explaining obvious things, restating the question
 - Hedging phrases: "I think maybe...", "It might be possible that..."
 - Hollow transitions: "Now, let's...", "Moving on to...", "Next, we'll..."
-- Colon headers: "**항목:** 설명", "**Topic:** content" - just write naturally without label-colon-content structure
+- Colon headers: "**Item:** description", "**Topic:** content" - just write naturally without label-colon-content structure
 
 **DO:**
 - Get to the point immediately
@@ -94,8 +94,8 @@ Utilize MCP servers whenever possible:
 - `firecrawl` - Web scraping and search
 
 ### Agent & Skill Usage
-- 작업에 적합한 agent 또는 skill이 있다면 적극 활용
-- 의도가 명확하면 바로 사용, 불명확하면 제안 후 승인 받기
+- Use appropriate agent or skill when available
+- If intent is clear, use directly; if unclear, propose and get approval
 
 ### Code Exploration Strategy
 
@@ -220,6 +220,10 @@ If pip install errors occur, try in order:
 - **Minimal Output**: Avoid unnecessary print statements in code - only include meaningful debug output when explicitly requested
 - **No Emojis**: Never use emojis anywhere, whether in documentation or code
 
+### Code Documentation Language
+- **Docstrings**: English (for API docs, IDE tooltips, type checker integration)
+- **Inline/block comments**: Korean (for quick understanding and clarity)
+
 ---
 
 ## File Organization

package/.claude/hooks/inject-guidelines.sh

@@ -10,11 +10,12 @@ GUIDELINES_DIR="$PROJECT_DIR/.claude/guidelines"
 
 # Primary guideline file to always inject
 PRIMARY_GUIDELINE="$GUIDELINES_DIR/work-guidelines.md"
+RELATIVE_GUIDELINE=".claude/guidelines/work-guidelines.md"
 
 # Output as system-reminder for Claude to process
 if [ -f "$PRIMARY_GUIDELINE" ]; then
   echo "<system-reminder>"
-  echo "Called the Read tool with the following input: {\"file_path\":\"$PRIMARY_GUIDELINE\"}"
+  echo "Called the Read tool with the following input: {\"file_path\":\"$RELATIVE_GUIDELINE\"}"
   echo "</system-reminder>"
   echo "<system-reminder>"
   echo "Result of calling the Read tool:"

package/.claude/hooks/notify_osc.sh

@@ -1,11 +1,11 @@
 #!/bin/bash
 # Claude Code OSC Notification Script
-# /dev/tty로 출력하여 터미널에 직접 전달
+# Output to /dev/tty to deliver directly to terminal
 
-# stdin에서 JSON 읽기 (timeout으로 blocking 방지)
+# Read JSON from stdin (prevent blocking with timeout)
 INPUT=$(timeout 1 cat 2>/dev/null || true)
 
-# JSON 파싱
+# JSON parsing
 if command -v jq &>/dev/null && [ -n "$INPUT" ]; then
     SESSION_ID=$(echo "$INPUT" | jq -r '.session_id // empty' 2>/dev/null)
     CWD=$(echo "$INPUT" | jq -r '.cwd // empty' 2>/dev/null)
@@ -16,12 +16,12 @@ if command -v jq &>/dev/null && [ -n "$INPUT" ]; then
     FOLDER=$(basename "${CWD:-$PWD}" 2>/dev/null)
     SHORT_ID="${SESSION_ID:0:8}"
 
-    # 제목 구성
+    # Title construction
     TITLE="Claude"
     [ -n "$FOLDER" ] && TITLE="$TITLE - $FOLDER"
     [ -n "$SHORT_ID" ] && TITLE="$TITLE [$SHORT_ID]"
 
-    # 본문 구성
+    # Message body construction
     case "$EVENT" in
         "Stop")
             BODY="Task completed"
@@ -42,8 +42,8 @@ else
     BODY="${2:-Task completed}"
 fi
 
-# OSC 777 알림 출력 (OSC 9는 중복 알림 발생하므로 제외)
-# /dev/tty 사용 가능 시 직접 출력, 아니면 stdout
+# OSC 777 notification output (excluding OSC 9 as it causes duplicate notifications)
+# Output directly to /dev/tty if available, otherwise stdout
 {
     printf '\033]777;notify;%s;%s\007' "$TITLE" "$BODY"
 } > /dev/tty 2>/dev/null || {

package/.claude/skills/code-explorer/SKILL.md

@@ -16,53 +16,53 @@ To search and analyze code repositories across GitHub and Hugging Face platforms
 | **GitHub** | Repositories, Code | `gh` CLI |
 | **Hugging Face** | Models, Datasets, Spaces | `uvx hf` CLI + `huggingface_hub` API |
 
-## 검색 품질 원칙 (필수)
+## Search Quality Principles (Required)
 
-### 1. 현재 날짜 확인
+### 1. Verify Current Date
 
 ```bash
 date +%Y-%m-%d
 ```
-아래 예시의 연도는 참고용. 실제 검색 시 **현재 연도** 사용.
+Years in examples below are for reference. Use **current year** in actual searches.
 
-### 2. Long-tail Keywords 적용
+### 2. Apply Long-tail Keywords
 
-예시의 Short-tail을 그대로 쓰지 말고 확장:
+Don't use short-tail examples as-is, expand them:
 
-| 예시 (참고용) | 실제 검색 (Long-tail) |
-|---------------|----------------------|
-| `"object detection"` | `"open vocabulary object detection pytorch inference {현재연도}"` |
+| Example (Reference) | Actual Search (Long-tail) |
+|---------------------|---------------------------|
+| `"object detection"` | `"open vocabulary object detection pytorch inference {current_year}"` |
 | `"gradio demo"` | `"gradio image segmentation interactive demo huggingface"` |
 | `"qwen vl"` | `"qwen2-vl vision language model zero-shot example code"` |
 
-### 3. Multi-Query 적용
+### 3. Apply Multi-Query
 
-한 번에 찾기 어려우면 **2-3개 관점**으로 검색:
+If hard to find at once, search from **2-3 perspectives**:
 
 ```bash
-# 모델명 중심
+# Model name focus
 gh search repos "qwen2-vl" --sort stars
 
-# 기능 중심
+# Feature focus
 gh search repos "vision language open vocabulary detection" --sort stars
 
-# 구현 중심
+# Implementation focus
 gh search repos "vl model gradio demo inference" --sort stars
 ```
 
-### 4. 필터 활용
+### 4. Use Filters
 
 ```bash
-# 최신 + 품질 필터 (현재 연도 적용)
-gh search repos "keyword" stars:>50 pushed:>{현재연도-1}-01-01 --language python
+# Recent + quality filter (apply current year)
+gh search repos "keyword" stars:>50 pushed:>{current_year-1}-01-01 --language python
 ```
 
-### 5. 검색 전 체크리스트
+### 5. Pre-Search Checklist
 
-- [ ] 현재 날짜 확인했는가?
-- [ ] Short-tail을 Long-tail로 변환했는가?
-- [ ] 필요시 2-3개 변형 쿼리로 검색했는가?
-- [ ] 적절한 필터(언어, 스타, 날짜) 적용했는가?
+- [ ] Verified current date?
+- [ ] Converted short-tail to long-tail?
+- [ ] Searched with 2-3 query variations if needed?
+- [ ] Applied appropriate filters (language, stars, date)?
 
 ---
 

package/.claude/skills/midjourney-imagineapi/SKILL.md

@@ -0,0 +1,260 @@
+---
+name: midjourney-imagineapi
+description: Generate optimized Midjourney V7 prompts and create images via ImagineAPI MCP server. Transform basic user requests into professional, cinematic prompts using 5-layer structure. Use when users request (1) Midjourney image generation, (2) prompt optimization for V7, (3) character/style reference usage, (4) professional photography prompts. Trigger phrases include "generate image", "create photo", "Midjourney", "make picture", "image generation".
+---
+
+# Midjourney Prompt Optimizer for ImagineAPI
+
+Transform user image requests into optimized Midjourney V7 prompts and generate images through ImagineAPI MCP.
+
+## Workflow Overview
+
+```
+User Request -> Clarify (if needed) -> Build Prompt -> Present Options -> Generate Image -> Return Results
+```
+
+## Step 1: Analyze User Request
+
+Determine what the user wants:
+- Main subject and concept
+- Implied style or mood
+- Any specific requirements mentioned
+
+If the request is clear and specific, proceed to Step 3.
+If the request is vague or generic (e.g., "make me a cat picture"), proceed to Step 2.
+
+## Step 2: Clarify with User (When Needed)
+
+For vague requests, use AskUserQuestion to clarify:
+
+**Style Clarification:**
+```
+AskUserQuestion:
+  header: "Style"
+  question: "What style do you want for this image?"
+  options:
+    - label: "Photorealistic"
+      description: "Looks like a real photograph"
+    - label: "Cinematic"
+      description: "Movie-like, dramatic lighting"
+    - label: "Artistic/Illustration"
+      description: "Stylized, artistic interpretation"
+    - label: "Minimal/Clean"
+      description: "Simple, clean aesthetic"
+```
+
+**Mood Clarification:**
+```
+AskUserQuestion:
+  header: "Mood"
+  question: "What mood or atmosphere?"
+  options:
+    - label: "Bright & Cheerful"
+      description: "Warm, optimistic feeling"
+    - label: "Moody & Dramatic"
+      description: "Dark, atmospheric, cinematic"
+    - label: "Calm & Serene"
+      description: "Peaceful, tranquil"
+    - label: "Energetic & Dynamic"
+      description: "Action, movement, excitement"
+```
+
+**Aspect Ratio:**
+```
+AskUserQuestion:
+  header: "Format"
+  question: "What format do you need?"
+  options:
+    - label: "Widescreen (16:9)"
+      description: "Cinematic, landscape, desktop"
+    - label: "Portrait (4:5)"
+      description: "Vertical, mobile, Instagram"
+    - label: "Square (1:1)"
+      description: "Social media, profile"
+    - label: "Ultra-wide (21:9)"
+      description: "Panoramic, banner"
+```
+
+## Step 3: Build Optimized Prompt
+
+Use the 5-layer structure from `references/midjourney-style-guide.md`:
+
+**Layer 1 (Required):** Subject & Composition
+```
+[Subject] [action/pose], [composition]
+```
+
+**Layer 2 (Required):** Style & Aesthetic
+```
+[style], [mood], [color treatment]
+```
+
+**Layer 3 (Recommended):** Lighting & Environment
+```
+[lighting], [time of day], [atmosphere]
+```
+
+**Layer 4 (Recommended):** Technical Details
+```
+[camera] [lens] [aperture], [depth of field], [quality terms]
+```
+
+**Layer 5 (Optional):** Artistic Reference
+```
+[artist/style reference]
+```
+
+Consult `references/prompt-examples.md` for transformation patterns.
+
+## Step 4: Select Parameters
+
+Based on the image type, select appropriate parameters from `references/midjourney-parameters.md`:
+
+**Portrait:**
+```
+--ar 4:5 --s 100-150 --q 2
+```
+
+**Landscape/Cinematic:**
+```
+--ar 16:9 --s 150-250 --q 2
+```
+
+**Product:**
+```
+--ar 1:1 --s 50-100
+```
+
+**Artistic:**
+```
+--ar varies --s 300-500 --exp 10-25
+```
+
+**Quick Iteration:**
+```
+--draft --c 25-50
+```
+
+## Step 5: Present Options to User
+
+Before generating, present 2-3 prompt variations using AskUserQuestion:
+
+```
+AskUserQuestion:
+  header: "Prompt"
+  question: "Which prompt direction do you prefer?"
+  options:
+    - label: "Option A: Dramatic"
+      description: "[show key differences - lighting, mood]"
+    - label: "Option B: Natural"
+      description: "[show key differences]"
+    - label: "Option C: Artistic"
+      description: "[show key differences]"
+```
+
+Show the full prompts in the response text, then ask for selection.
+
+## Step 6: Generate Image
+
+Use the midjourney-proxy-mcp MCP tools directly. See `references/imagineapi-integration.md` for details.
+
+**Generate:**
+Call the `midjourney_imagine` MCP tool with the optimized prompt:
+- `prompt`: The complete prompt string including all parameters
+
+**Check Status:**
+Call the `midjourney_get_status` MCP tool:
+- `image_id`: The ID returned from midjourney_imagine
+
+Poll every 10-15 seconds until status is "completed".
+
+## Step 7: Return Results
+
+Provide to user:
+1. Generated image URL
+2. Upscaled versions (if available in upscaled_urls)
+3. The prompt used (for user's reference)
+4. Offer to create variations if needed
+
+---
+
+## Reference Files
+
+### midjourney-parameters.md
+Complete V7 parameter reference. Consult when:
+- Selecting appropriate parameters
+- Explaining parameter effects to user
+- Troubleshooting parameter combinations
+
+Search patterns:
+- "Quick Reference Table" - All parameters at a glance
+- "V7-Specific Parameters" - Draft, exp, oref
+- "Reference System" - Style and character references
+
+### midjourney-style-guide.md
+Prompt structure and vocabulary. Consult when:
+- Building prompts from scratch
+- Selecting descriptive vocabulary
+- Checking quality guidelines
+
+Search patterns:
+- "5-Layer Prompt Structure" - Building prompts
+- "Descriptive Vocabulary" - Lighting, mood, color terms
+- "Genre-Specific Patterns" - Templates by category
+
+### prompt-examples.md
+Before/after transformations. Consult when:
+- Learning transformation patterns
+- Finding inspiration for specific genres
+- Understanding enhancement techniques
+
+Search patterns:
+- "Portrait Photography" - People, characters
+- "Landscape/Environment" - Nature, cities
+- "Product Photography" - Objects, commercial
+- "Character Consistency" - Using references
+
+### imagineapi-integration.md
+MCP tool usage. Consult when:
+- Calling imagine tool
+- Checking generation status
+- Handling errors
+
+---
+
+## Quick Reference
+
+### Minimum Viable Prompt
+```
+[subject], [style], [lighting], [camera spec] --ar [ratio] --s [value]
+```
+
+### Full Prompt Template
+```
+[Subject description], [composition], [style category], [lighting type],
+[mood/atmosphere], shot on [camera] [lens], [color treatment],
+[quality markers] --ar [ratio] --s [value] --q [value]
+```
+
+### Common Parameter Combos
+
+| Purpose | Parameters |
+|---------|------------|
+| Quick test | `--draft` |
+| Standard quality | `--ar 16:9 --s 100` |
+| High quality | `--ar 16:9 --s 150 --q 2` |
+| Artistic | `--s 400 --exp 25` |
+| Character lock | `--oref <url> --ow 200` |
+| Style lock | `--sref <url> --sw 300` |
+| Both locks | `--oref <url> --ow 150 --sref <url> --sw 200` |
+
+---
+
+## Important Notes
+
+1. V7 does not support multi-prompting (`::` syntax)
+2. Use natural language - V7 understands creative intent
+3. Keep prompts 20-75 words for best results
+4. Always append parameters after the descriptive text
+5. Use --draft for rapid iteration, then enhance winners
+6. Combine --oref and --sref for maximum consistency