feat: v3.2 나머지 미완성 기능 구현

[Instagram Reels] Phase 2 완성
- instagram_bot.py: publish_reels() 추가 (MP4 → Reels API)
  - upload_video_container(), wait_for_video_ready() 구현
  - 로컬 경로 → 공개 URL 자동 변환 (image_host.get_public_video_url())
- scheduler.py: job_distribute_instagram_reels() 추가 (10:30)
- image_host.py: get_public_video_url() + 로컬 비디오 서버 추가
  - VIDEO_HOST_BASE_URL 환경변수 지원 (Tailscale/CDN)

[writer_bot.py] 신규 — 독립 실행형 글쓰기 봇
- api_content.py manual-write 엔드포인트에서 subprocess 호출 가능
- run_pending(): 오늘 날짜 미처리 글감 자동 처리
- run_from_topic(): 직접 주제 지정
- run_from_file(): JSON 파일 지정
- CLI: python bots/writer_bot.py [--topic "..." | --file path.json | --limit N]

[보조 시스템 신규] v3.1 CLI + Assist 모드
- blog.cmd: venv Python 경유 Windows 런처
- blog_runtime.py + runtime_guard.py: 실행 진입점 + venv 검증
- blog_engine_cli.py: 대시보드 API 기반 CLI (blog status, blog review 등)
- bots/assist_bot.py: URL 기반 수동 어시스트 파이프라인
- dashboard/backend/api_assist.py + frontend/Assist.jsx: 수동모드 탭

[engine_loader.py] v3.1 개선
- OpenClawWriter: --json 플래그 + payloads 파싱 + plain text 폴백
- ClaudeWebWriter: Playwright 쿠키 세션 (Cloudflare 차단으로 현재 비활성)
- GeminiWebWriter: gemini-webapi 비공식 클라이언트

[scheduler.py] v3.1 개선
- _call_openclaw(): 플레이스홀더 → EngineLoader 실제 호출
- _build_openclaw_prompt(): 구조화된 HTML 원고 프롬프트
- data/originals/: 원본 article JSON 저장 경로 추가

[설정/환경] 정비
- .env.example: SEEDANCE/ELEVENLABS/GEMINI/RUNWAY 복원
  + VIDEO_HOST_BASE_URL, GEMINI_WEB_* , REMOTE_CLAUDE_POLLING_ENABLED 추가
- scripts/setup.bat: data/originals, outputs, assist, novels, config/novels
  디렉토리 생성 + 폰트 다운로드 + blog.cmd 기반 Task Scheduler 등록
- requirements.txt: fastapi, uvicorn, python-multipart 추가

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

scripts/setup.bat

@@ -36,16 +36,27 @@ if not exist data\published mkdir data\published
 if not exist data\analytics mkdir data\analytics
 if not exist data\images mkdir data\images
 if not exist data\drafts mkdir data\drafts
+if not exist data\originals mkdir data\originals
+if not exist data\outputs mkdir data\outputs
+if not exist data\assist mkdir data\assist
+if not exist data\assist\sessions mkdir data\assist\sessions
+if not exist data\assist\inbox mkdir data\assist\inbox
+if not exist data\novels mkdir data\novels
 if not exist logs mkdir logs
+if not exist config\novels mkdir config\novels
 
-REM Register scheduler.py in Windows Task Scheduler
-set SCRIPT_PATH=%~dp0bots\scheduler.py
+REM Download fonts (Noto Sans KR for card/shorts converter)
+echo [INFO] Downloading fonts...
+venv\Scripts\python.exe scripts\download_fonts.py
+
+REM Register scheduler.py in Windows Task Scheduler (blog.cmd 경유)
+set BLOG_CMD=%~dp0blog.cmd
 set PYTHON_PATH=%~dp0venv\Scripts\pythonw.exe
 
 schtasks /query /tn "BlogEngine" >nul 2>&1
 if errorlevel 1 (
-    schtasks /create /tn "BlogEngine" /tr "\"%PYTHON_PATH%\" \"%SCRIPT_PATH%\"" /sc onlogon /rl highest /f
-    echo [OK] BlogEngine registered in Windows Task Scheduler
+    schtasks /create /tn "BlogEngine" /tr "\"%BLOG_CMD%\" scheduler" /sc onlogon /rl highest /f
+    echo [OK] BlogEngine registered in Windows Task Scheduler (blog scheduler)
 ) else (
     echo [INFO] BlogEngine task already registered.
 )
@@ -58,7 +69,10 @@ echo.
 echo Next steps:
 echo 1. Open .env and fill in all API keys
 echo 2. Run scripts\get_token.py to get Google OAuth token
-echo 3. Update BLOG_MAIN_ID in config\blogs.json with your actual blog ID
-echo 4. Start scheduler with: python bots\scheduler.py
+echo    (Blogger + Search Console + YouTube OAuth)
+echo 3. Update BLOG_MAIN_ID in .env with your Blogger blog ID
+echo 4. Start scheduler:  blog scheduler
+echo 5. Start dashboard:  blog server  (http://localhost:8080)
+echo 6. CLI status check: blog status
 echo.
 pause