feat: 전체 PKM 스크립트 일괄 작성 — 분류/법령/메일/다이제스트/임베딩

- scripts/pkm_utils.py: 공통 유틸 (로거, dotenv, osascript 래퍼)
- scripts/prompts/classify_document.txt: Ollama 분류 프롬프트
- applescript/auto_classify.scpt: Inbox → AI 분류 → DB 이동
- applescript/omnifocus_sync.scpt: Projects → OmniFocus 작업 생성
- scripts/law_monitor.py: 법령 변경 모니터링 + DEVONthink 임포트
- scripts/mailplus_archive.py: MailPlus IMAP → Archive DB
- scripts/pkm_daily_digest.py: 일일 다이제스트 + OmniFocus 액션
- scripts/embed_to_chroma.py: GPU 서버 벡터 임베딩 → ChromaDB
- launchd/*.plist: 3개 스케줄 (07:00, 07:00+18:00, 20:00)
- docs/deploy.md: Mac mini 배포 가이드
- docs/devonagent-setup.md: 검색 세트 9종 설정 가이드
- tests/test_classify.py: 5종 문서 분류 테스트

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

applescript/auto_classify.scpt

@@ -0,0 +1,101 @@
+-- DEVONthink 4 Smart Rule: AI 자동 분류
+-- Inbox DB 새 문서 → Ollama 분류 → 태그 + 메타데이터 + 도메인 DB 이동
+-- Smart Rule 설정: Event = On Import, 조건 = Tags is empty
+
+on performSmartRule(theRecords)
+	tell application id "DNtp"
+		repeat with theRecord in theRecords
+			try
+				-- 1. 문서 텍스트 추출 (최대 4000자)
+				set docText to plain text of theRecord
+				set docUUID to uuid of theRecord
+
+				if length of docText > 4000 then
+					set docText to text 1 thru 4000 of docText
+				end if
+
+				if length of docText < 10 then
+					-- 텍스트가 너무 짧으면 건너뜀
+					set tags of theRecord to {"@상태/검토필요"}
+					continue repeat
+				end if
+
+				-- 2. 분류 프롬프트 로딩
+				set promptPath to (POSIX path of (path to home folder)) & "Documents/code/DEVONThink_my server/scripts/prompts/classify_document.txt"
+				set promptTemplate to do shell script "cat " & quoted form of promptPath
+
+				-- 문서 텍스트를 프롬프트에 삽입 (특수문자 이스케이프)
+				set escapedText to do shell script "echo " & quoted form of docText & " | sed 's/\\\\/\\\\\\\\/g; s/\"/\\\\\"/g; s/\\n/\\\\n/g' | head -c 4000"
+
+				-- 3. Ollama API 호출
+				set curlCmd to "curl -s --max-time 120 http://localhost:11434/api/generate -d '{\"model\": \"qwen3.5:35b-a3b-q4_K_M\", \"prompt\": " & quoted form of escapedText & ", \"stream\": false, \"format\": \"json\"}'"
+				set jsonResult to do shell script curlCmd
+
+				-- 4. JSON 파싱 (Python 사용)
+				set parseCmd to "echo " & quoted form of jsonResult & " | python3 -c \"
+import sys, json
+try:
+    r = json.loads(sys.stdin.read())
+    d = json.loads(r.get('response', '{}'))
+    tags = ','.join(d.get('tags', []))
+    db = d.get('domain_db', '00_Note_BOX')
+    grp = d.get('sub_group', '00_Inbox')
+    ch = d.get('sourceChannel', 'inbox_route')
+    origin = d.get('dataOrigin', 'external')
+    print(f'{db}|{grp}|{tags}|{ch}|{origin}')
+except:
+    print('00_Note_BOX|00_Inbox||inbox_route|external')
+\""
+
+				set classResult to do shell script parseCmd
+				set AppleScript's text item delimiters to "|"
+				set resultParts to text items of classResult
+				set targetDB to item 1 of resultParts
+				set targetGroup to item 2 of resultParts
+				set tagString to item 3 of resultParts
+				set sourceChannel to item 4 of resultParts
+				set dataOrigin to item 5 of resultParts
+				set AppleScript's text item delimiters to ""
+
+				-- 5. 태그 설정
+				if tagString is not "" then
+					set AppleScript's text item delimiters to ","
+					set tagList to text items of tagString
+					set AppleScript's text item delimiters to ""
+					set tags of theRecord to tagList
+				end if
+
+				-- 6. 커스텀 메타데이터 설정
+				add custom meta data sourceChannel for "sourceChannel" to theRecord
+				add custom meta data dataOrigin for "dataOrigin" to theRecord
+				add custom meta data (current date) for "lastAIProcess" to theRecord
+				add custom meta data "inbox_route" for "sourceChannel" to theRecord
+
+				-- 7. 대상 도메인 DB로 이동
+				set targetDatabase to missing value
+				repeat with db in databases
+					if name of db is targetDB then
+						set targetDatabase to db
+						exit repeat
+					end if
+				end repeat
+
+				if targetDatabase is not missing value then
+					set groupPath to "/" & targetGroup
+					set targetLocation to create location groupPath in targetDatabase
+					move record theRecord to targetLocation
+				end if
+
+				-- 8. GPU 서버 벡터 임베딩 비동기 전송
+				set embedScript to (POSIX path of (path to home folder)) & "Documents/code/DEVONThink_my server/venv/bin/python3"
+				set embedPy to (POSIX path of (path to home folder)) & "Documents/code/DEVONThink_my server/scripts/embed_to_chroma.py"
+				do shell script embedScript & " " & quoted form of embedPy & " " & quoted form of docUUID & " &> /dev/null &"
+
+			on error errMsg
+				-- 에러 시 로그 기록 + 검토필요 태그
+				set tags of theRecord to {"@상태/검토필요", "AI분류실패"}
+				do shell script "echo '[" & (current date) & "] [auto_classify] [ERROR] " & errMsg & "' >> ~/Documents/code/DEVONThink_my\\ server/logs/auto_classify.log"
+			end try
+		end repeat
+	end tell
+end performSmartRule

applescript/omnifocus_sync.scpt

@@ -0,0 +1,71 @@
+-- DEVONthink 4 Smart Rule: OmniFocus 연동
+-- Projects DB 새 문서에서 TODO 패턴 감지 → OmniFocus 작업 생성
+-- Smart Rule 설정: Event = On Import, DB = Projects
+
+on performSmartRule(theRecords)
+	tell application id "DNtp"
+		repeat with theRecord in theRecords
+			try
+				set docText to plain text of theRecord
+				set docTitle to name of theRecord
+				set docUUID to uuid of theRecord
+				set docLink to reference URL of theRecord -- x-devonthink-item://UUID
+
+				-- TODO 패턴 감지: "TODO", "할일", "□", "[ ]", "FIXME"
+				set hasAction to false
+				if docText contains "TODO" or docText contains "할일" or docText contains "□" or docText contains "[ ]" or docText contains "FIXME" then
+					set hasAction to true
+				end if
+
+				if not hasAction then continue repeat
+
+				-- 액션 아이템 추출 (Python으로 파싱)
+				set extractCmd to "echo " & quoted form of docText & " | python3 -c \"
+import sys, re
+text = sys.stdin.read()
+patterns = [
+    r'(?:TODO|FIXME|할일)[:\\s]*(.+?)(?:\\n|$)',
+    r'(?:□|\\[ \\])\\s*(.+?)(?:\\n|$)',
+]
+items = []
+for p in patterns:
+    items.extend(re.findall(p, text, re.MULTILINE))
+# 최대 5개, 중복 제거
+seen = set()
+for item in items[:10]:
+    item = item.strip()
+    if item and item not in seen:
+        seen.add(item)
+        print(item)
+    if len(seen) >= 5:
+        break
+\""
+
+				set actionItems to paragraphs of (do shell script extractCmd)
+
+				if (count of actionItems) = 0 then continue repeat
+
+				-- OmniFocus에 작업 생성
+				tell application "OmniFocus"
+					tell default document
+						set taskIDs to {}
+						repeat with actionItem in actionItems
+							set taskName to docTitle & " — " & (contents of actionItem)
+							set newTask to make new inbox task with properties {name:taskName, note:"DEVONthink 문서: " & docLink}
+							set end of taskIDs to id of newTask
+						end repeat
+					end tell
+				end tell
+
+				-- DEVONthink 메타데이터에 OmniFocus Task ID 저장
+				set AppleScript's text item delimiters to ","
+				set taskIDString to taskIDs as text
+				set AppleScript's text item delimiters to ""
+				add custom meta data taskIDString for "omnifocusTaskID" to theRecord
+
+			on error errMsg
+				do shell script "echo '[" & (current date) & "] [omnifocus_sync] [ERROR] " & errMsg & "' >> ~/Documents/code/DEVONThink_my\\ server/logs/omnifocus_sync.log"
+			end try
+		end repeat
+	end tell
+end performSmartRule

docs/deploy.md

@@ -0,0 +1,96 @@
+# Mac mini 배포 가이드
+
+## 1. 초기 설치
+
+```bash
+# Mac mini에서
+cd ~/Documents/code/
+git clone https://git.hyungi.net/hyungi/devonthink_home.git "DEVONThink_my server"
+cd "DEVONThink_my server"
+
+# Python 가상환경
+python3 -m venv venv
+source venv/bin/activate
+pip install -r requirements.txt
+```
+
+## 2. 인증 정보 설정
+
+```bash
+mkdir -p ~/.config/pkm
+nano ~/.config/pkm/credentials.env
+chmod 600 ~/.config/pkm/credentials.env
+```
+
+credentials.env.example을 참고하여 실제 값 입력.
+
+## 3. launchd 스케줄 등록
+
+```bash
+# 심볼릭 링크 생성
+ln -sf ~/Documents/code/DEVONThink_my\ server/launchd/net.hyungi.pkm.law-monitor.plist ~/Library/LaunchAgents/
+ln -sf ~/Documents/code/DEVONThink_my\ server/launchd/net.hyungi.pkm.mailplus.plist ~/Library/LaunchAgents/
+ln -sf ~/Documents/code/DEVONThink_my\ server/launchd/net.hyungi.pkm.daily-digest.plist ~/Library/LaunchAgents/
+
+# 등록
+launchctl load ~/Library/LaunchAgents/net.hyungi.pkm.law-monitor.plist
+launchctl load ~/Library/LaunchAgents/net.hyungi.pkm.mailplus.plist
+launchctl load ~/Library/LaunchAgents/net.hyungi.pkm.daily-digest.plist
+
+# 확인
+launchctl list | grep pkm
+```
+
+## 4. 수동 테스트
+
+```bash
+cd ~/Documents/code/DEVONThink_my\ server/
+source venv/bin/activate
+
+# 각 스크립트 수동 실행
+python3 scripts/law_monitor.py
+python3 scripts/mailplus_archive.py
+python3 scripts/pkm_daily_digest.py
+```
+
+## 5. DEVONthink Smart Rule 설정
+
+1. DEVONthink → Preferences → Smart Rules
+2. 새 Rule: "AI Auto Classify"
+   - Event: On Import
+   - Database: Inbox
+   - Condition: Tags is empty
+   - Action: Execute Script → External → `applescript/auto_classify.scpt`
+3. 새 Rule: "OmniFocus Sync"
+   - Event: On Import
+   - Database: Projects
+   - Action: Execute Script → External → `applescript/omnifocus_sync.scpt`
+
+## 6. 업데이트
+
+```bash
+cd ~/Documents/code/DEVONThink_my\ server/
+git pull
+source venv/bin/activate
+pip install -r requirements.txt
+```
+
+## 7. 로그 확인
+
+```bash
+# 스크립트 로그
+tail -f logs/law_monitor.log
+tail -f logs/mailplus.log
+tail -f logs/digest.log
+
+# launchd 로그
+tail -f logs/law_monitor_launchd.log
+```
+
+## 실행 스케줄
+
+| 스크립트 | 시간 | 용도 |
+|---------|------|------|
+| law_monitor.py | 매일 07:00 | 법령 변경 모니터링 |
+| mailplus_archive.py | 매일 07:00, 18:00 | 이메일 수집 |
+| pkm_daily_digest.py | 매일 20:00 | 일일 다이제스트 |

docs/devonagent-setup.md

@@ -0,0 +1,103 @@
+# DEVONagent Pro — 검색 세트 설정 가이드
+
+DEVONagent Pro에서 안전 분야 + 기술 분야 자동 검색 세트를 설정합니다.
+주간 합계 50~85건 수준으로 양을 조절합니다.
+
+## 공통 설정
+
+- **Schedule**: 각 세트별 지정 (매일/주간)
+- **Action**: Import to DEVONthink → Inbox DB
+- **Max Results per Run**: 각 세트별 지정
+- **Language**: 해당 언어
+
+---
+
+## 검색 세트 1: 국내 산업안전 뉴스 (매일)
+
+- **키워드**: `산업안전 OR 중대재해 OR 위험성평가 OR 안전사고`
+- **사이트**: kosha.or.kr, moel.go.kr, safetynews.co.kr, dailysafety.com
+- **Max Results**: 5/일
+- **Schedule**: 매일 08:00
+
+## 검색 세트 2: 국내 중대재해 뉴스 (매일)
+
+- **키워드**: `중대재해 OR 산업재해 OR 작업장사고 -주식 -부동산`
+- **사이트**: 뉴스 전체
+- **Max Results**: 3/일
+- **Schedule**: 매일 08:30
+
+## 검색 세트 3: KOSHA 가이드/지침 (주간)
+
+- **키워드**: `site:kosha.or.kr 가이드 OR 지침 OR 기술자료`
+- **Max Results**: 5/주
+- **Schedule**: 매주 월요일 09:00
+
+## 검색 세트 4: 국내 산업안전 학술/논문 (주간)
+
+- **키워드**: `산업안전 OR 위험성평가 OR occupational safety site:kci.go.kr OR site:dbpia.co.kr`
+- **Max Results**: 3/주
+- **Schedule**: 매주 수요일 09:00
+
+## 검색 세트 5: US OSHA / Safety+Health Magazine (주간)
+
+- **키워드**: `occupational safety OR workplace hazard OR OSHA regulation`
+- **사이트**: osha.gov, safetyandhealthmagazine.com, ehstoday.com
+- **Max Results**: 5/주
+- **Language**: English
+- **Schedule**: 매주 화요일 09:00
+
+## 검색 세트 6: JP 厚生労働省 / 安全衛生 (주간)
+
+- **키워드**: `労働安全 OR 安全衛生 OR 労災`
+- **사이트**: mhlw.go.jp, jisha.or.jp
+- **Max Results**: 3/주
+- **Language**: Japanese
+- **Schedule**: 매주 목요일 09:00
+
+## 검색 세트 7: EU-OSHA (월간)
+
+- **키워드**: `occupational safety health EU regulation`
+- **사이트**: osha.europa.eu
+- **Max Results**: 5/월
+- **Language**: English
+- **Schedule**: 매월 1일 09:00
+
+## 검색 세트 8: 기술 뉴스 — AI/서버/네트워크 (매일)
+
+- **키워드**: `AI model release OR server infrastructure OR homelab OR self-hosted`
+- **사이트**: news.ycombinator.com, arstechnica.com, theregister.com
+- **Max Results**: 5/일
+- **Schedule**: 매일 12:00
+
+## 검색 세트 9: 프로그래밍 기술 동향 (주간)
+
+- **키워드**: `Python release OR Node.js update OR Docker best practice OR FastAPI`
+- **사이트**: dev.to, blog.python.org, nodejs.org
+- **Max Results**: 5/주
+- **Schedule**: 매주 금요일 12:00
+
+---
+
+## 주간 예상 건수
+
+| 세트 | 빈도 | 건/주 |
+|------|------|-------|
+| 1. 국내 안전 뉴스 | 매일 5 | ~35 |
+| 2. 중대재해 뉴스 | 매일 3 | ~21 |
+| 3. KOSHA 가이드 | 주간 5 | 5 |
+| 4. 학술/논문 | 주간 3 | 3 |
+| 5. US OSHA | 주간 5 | 5 |
+| 6. JP 안전위생 | 주간 3 | 3 |
+| 7. EU-OSHA | 월간 5 | ~1 |
+| 8. 기술 뉴스 | 매일 5 | ~35 |
+| 9. 프로그래밍 | 주간 5 | 5 |
+| **합계** | | **~113** |
+
+> 양이 너무 많으면 세트 1, 2, 8의 Max Results를 3으로 줄이면 주간 ~65건 수준으로 조절 가능.
+
+## DEVONthink 전송 설정
+
+1. DEVONagent → Preferences → DEVONthink
+2. Target Database: **Inbox**
+3. Auto-Tag: 검색 세트 이름으로 자동 태그 (`devonagent-검색세트명`)
+4. DEVONthink Smart Rule이 Inbox에서 자동 분류 처리

launchd/net.hyungi.pkm.daily-digest.plist

@@ -0,0 +1,28 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+    <key>Label</key>
+    <string>net.hyungi.pkm.daily-digest</string>
+    <key>ProgramArguments</key>
+    <array>
+        <string>/Users/hyungi/Documents/code/DEVONThink_my server/venv/bin/python3</string>
+        <string>/Users/hyungi/Documents/code/DEVONThink_my server/scripts/pkm_daily_digest.py</string>
+    </array>
+    <key>WorkingDirectory</key>
+    <string>/Users/hyungi/Documents/code/DEVONThink_my server</string>
+    <key>StartCalendarInterval</key>
+    <dict>
+        <key>Hour</key>
+        <integer>20</integer>
+        <key>Minute</key>
+        <integer>0</integer>
+    </dict>
+    <key>StandardOutPath</key>
+    <string>/Users/hyungi/Documents/code/DEVONThink_my server/logs/digest_launchd.log</string>
+    <key>StandardErrorPath</key>
+    <string>/Users/hyungi/Documents/code/DEVONThink_my server/logs/digest_launchd_err.log</string>
+    <key>RunAtLoad</key>
+    <false/>
+</dict>
+</plist>

launchd/net.hyungi.pkm.law-monitor.plist

@@ -0,0 +1,28 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+    <key>Label</key>
+    <string>net.hyungi.pkm.law-monitor</string>
+    <key>ProgramArguments</key>
+    <array>
+        <string>/Users/hyungi/Documents/code/DEVONThink_my server/venv/bin/python3</string>
+        <string>/Users/hyungi/Documents/code/DEVONThink_my server/scripts/law_monitor.py</string>
+    </array>
+    <key>WorkingDirectory</key>
+    <string>/Users/hyungi/Documents/code/DEVONThink_my server</string>
+    <key>StartCalendarInterval</key>
+    <dict>
+        <key>Hour</key>
+        <integer>7</integer>
+        <key>Minute</key>
+        <integer>0</integer>
+    </dict>
+    <key>StandardOutPath</key>
+    <string>/Users/hyungi/Documents/code/DEVONThink_my server/logs/law_monitor_launchd.log</string>
+    <key>StandardErrorPath</key>
+    <string>/Users/hyungi/Documents/code/DEVONThink_my server/logs/law_monitor_launchd_err.log</string>
+    <key>RunAtLoad</key>
+    <false/>
+</dict>
+</plist>

launchd/net.hyungi.pkm.mailplus.plist

@@ -0,0 +1,36 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+    <key>Label</key>
+    <string>net.hyungi.pkm.mailplus</string>
+    <key>ProgramArguments</key>
+    <array>
+        <string>/Users/hyungi/Documents/code/DEVONThink_my server/venv/bin/python3</string>
+        <string>/Users/hyungi/Documents/code/DEVONThink_my server/scripts/mailplus_archive.py</string>
+    </array>
+    <key>WorkingDirectory</key>
+    <string>/Users/hyungi/Documents/code/DEVONThink_my server</string>
+    <key>StartCalendarInterval</key>
+    <array>
+        <dict>
+            <key>Hour</key>
+            <integer>7</integer>
+            <key>Minute</key>
+            <integer>0</integer>
+        </dict>
+        <dict>
+            <key>Hour</key>
+            <integer>18</integer>
+            <key>Minute</key>
+            <integer>0</integer>
+        </dict>
+    </array>
+    <key>StandardOutPath</key>
+    <string>/Users/hyungi/Documents/code/DEVONThink_my server/logs/mailplus_launchd.log</string>
+    <key>StandardErrorPath</key>
+    <string>/Users/hyungi/Documents/code/DEVONThink_my server/logs/mailplus_launchd_err.log</string>
+    <key>RunAtLoad</key>
+    <false/>
+</dict>
+</plist>

scripts/embed_to_chroma.py

@@ -0,0 +1,104 @@
+#!/usr/bin/env python3
+"""
+벡터 임베딩 스크립트
+- DEVONthink 문서 UUID로 텍스트 추출
+- GPU 서버(nomic-embed-text)로 임베딩 생성
+- ChromaDB에 저장
+"""
+
+import os
+import sys
+import requests
+from pathlib import Path
+
+sys.path.insert(0, str(Path(__file__).parent))
+from pkm_utils import setup_logger, load_credentials, run_applescript_inline
+
+logger = setup_logger("embed")
+
+# ChromaDB 저장 경로
+CHROMA_DIR = Path.home() / ".local" / "share" / "pkm" / "chromadb"
+CHROMA_DIR.mkdir(parents=True, exist_ok=True)
+
+
+def get_document_text(uuid: str) -> tuple[str, str]:
+    """DEVONthink에서 UUID로 문서 텍스트 + 제목 추출"""
+    script = f'''
+    tell application id "DNtp"
+        set theRecord to get record with uuid "{uuid}"
+        set docText to plain text of theRecord
+        set docTitle to name of theRecord
+        return docTitle & "|||" & docText
+    end tell
+    '''
+    result = run_applescript_inline(script)
+    parts = result.split("|||", 1)
+    title = parts[0] if len(parts) > 0 else ""
+    text = parts[1] if len(parts) > 1 else ""
+    return title, text
+
+
+def get_embedding(text: str, gpu_server_ip: str) -> list[float] | None:
+    """GPU 서버의 nomic-embed-text로 임베딩 생성"""
+    url = f"http://{gpu_server_ip}:11434/api/embeddings"
+    try:
+        resp = requests.post(url, json={
+            "model": "nomic-embed-text",
+            "prompt": text[:8000]  # 토큰 제한
+        }, timeout=60)
+        resp.raise_for_status()
+        return resp.json().get("embedding")
+    except Exception as e:
+        logger.error(f"임베딩 생성 실패: {e}")
+        return None
+
+
+def store_in_chromadb(doc_id: str, title: str, text: str, embedding: list[float]):
+    """ChromaDB에 저장"""
+    import chromadb
+    client = chromadb.PersistentClient(path=str(CHROMA_DIR))
+    collection = client.get_or_create_collection(
+        name="pkm_documents",
+        metadata={"hnsw:space": "cosine"}
+    )
+    collection.upsert(
+        ids=[doc_id],
+        embeddings=[embedding],
+        documents=[text[:2000]],
+        metadatas=[{"title": title, "source": "devonthink"}]
+    )
+    logger.info(f"ChromaDB 저장: {doc_id} ({title[:30]})")
+
+
+def run(uuid: str):
+    """단일 문서 임베딩 처리"""
+    logger.info(f"임베딩 처리 시작: {uuid}")
+
+    creds = load_credentials()
+    gpu_ip = creds.get("GPU_SERVER_IP")
+    if not gpu_ip:
+        logger.warning("GPU_SERVER_IP 미설정 — 임베딩 건너뜀")
+        return
+
+    try:
+        title, text = get_document_text(uuid)
+        if not text or len(text) < 10:
+            logger.warning(f"텍스트 부족 [{uuid}]: {len(text)}자")
+            return
+
+        embedding = get_embedding(text, gpu_ip)
+        if embedding:
+            store_in_chromadb(uuid, title, text, embedding)
+            logger.info(f"임베딩 완료: {uuid}")
+        else:
+            logger.error(f"임베딩 실패: {uuid}")
+
+    except Exception as e:
+        logger.error(f"임베딩 처리 에러 [{uuid}]: {e}", exc_info=True)
+
+
+if __name__ == "__main__":
+    if len(sys.argv) < 2:
+        print("사용법: python3 embed_to_chroma.py <DEVONthink_UUID>")
+        sys.exit(1)
+    run(sys.argv[1])

scripts/law_monitor.py

@@ -0,0 +1,184 @@
+#!/usr/bin/env python3
+"""
+법령 모니터링 스크립트
+- 국가법령정보센터 OpenAPI (open.law.go.kr) 폴링
+- 산업안전보건법, 중대재해처벌법 등 변경 추적
+- 변경 감지 시 DEVONthink 04_Industrial Safety 자동 임포트
+※ API 승인 대기중 — 스크립트만 작성, 실제 호출은 승인 후
+"""
+
+import os
+import sys
+import json
+import requests
+import xml.etree.ElementTree as ET
+from datetime import datetime, timedelta
+from pathlib import Path
+
+sys.path.insert(0, str(Path(__file__).parent))
+from pkm_utils import setup_logger, load_credentials, run_applescript_inline, PROJECT_ROOT, DATA_DIR
+
+logger = setup_logger("law_monitor")
+
+# 모니터링 대상 법령
+MONITORED_LAWS = [
+    {"name": "산업안전보건법", "law_id": "001789", "category": "법률"},
+    {"name": "산업안전보건법 시행령", "law_id": "001790", "category": "대통령령"},
+    {"name": "산업안전보건법 시행규칙", "law_id": "001791", "category": "부령"},
+    {"name": "중대재해 처벌 등에 관한 법률", "law_id": "019005", "category": "법률"},
+    {"name": "중대재해 처벌 등에 관한 법률 시행령", "law_id": "019006", "category": "대통령령"},
+    {"name": "화학물질관리법", "law_id": "012354", "category": "법률"},
+    {"name": "위험물안전관리법", "law_id": "001478", "category": "법률"},
+]
+
+# 마지막 확인 일자 저장 파일
+LAST_CHECK_FILE = DATA_DIR / "law_last_check.json"
+LAWS_DIR = DATA_DIR / "laws"
+LAWS_DIR.mkdir(exist_ok=True)
+
+
+def load_last_check() -> dict:
+    """마지막 확인 일자 로딩"""
+    if LAST_CHECK_FILE.exists():
+        with open(LAST_CHECK_FILE, "r") as f:
+            return json.load(f)
+    return {}
+
+
+def save_last_check(data: dict):
+    """마지막 확인 일자 저장"""
+    with open(LAST_CHECK_FILE, "w") as f:
+        json.dump(data, f, ensure_ascii=False, indent=2)
+
+
+def fetch_law_info(law_oc: str, law_id: str) -> dict | None:
+    """법령 정보 조회 (법령 API)"""
+    url = "https://www.law.go.kr/DRF/lawSearch.do"
+    params = {
+        "OC": law_oc,
+        "target": "law",
+        "type": "JSON",
+        "MST": law_id,
+    }
+    try:
+        resp = requests.get(url, params=params, timeout=30)
+        resp.raise_for_status()
+        data = resp.json()
+        if "LawSearch" in data and "law" in data["LawSearch"]:
+            laws = data["LawSearch"]["law"]
+            if isinstance(laws, list):
+                return laws[0] if laws else None
+            return laws
+        return None
+    except Exception as e:
+        logger.error(f"법령 조회 실패 [{law_id}]: {e}")
+        return None
+
+
+def fetch_law_text(law_oc: str, law_mst: str) -> str | None:
+    """법령 본문 XML 다운로드"""
+    url = "https://www.law.go.kr/DRF/lawService.do"
+    params = {
+        "OC": law_oc,
+        "target": "law",
+        "type": "XML",
+        "MST": law_mst,
+    }
+    try:
+        resp = requests.get(url, params=params, timeout=60)
+        resp.raise_for_status()
+        return resp.text
+    except Exception as e:
+        logger.error(f"법령 본문 다운로드 실패 [{law_mst}]: {e}")
+        return None
+
+
+def save_law_file(law_name: str, content: str) -> Path:
+    """법령 XML 저장"""
+    today = datetime.now().strftime("%Y%m%d")
+    safe_name = law_name.replace(" ", "_").replace("/", "_")
+    filepath = LAWS_DIR / f"{safe_name}_{today}.xml"
+    with open(filepath, "w", encoding="utf-8") as f:
+        f.write(content)
+    logger.info(f"법령 저장: {filepath}")
+    return filepath
+
+
+def import_to_devonthink(filepath: Path, law_name: str, category: str):
+    """DEVONthink 04_Industrial Safety로 임포트"""
+    script = f'''
+    tell application id "DNtp"
+        set targetDB to missing value
+        repeat with db in databases
+            if name of db is "04_Industrial safety" then
+                set targetDB to db
+                exit repeat
+            end if
+        end repeat
+
+        if targetDB is not missing value then
+            set targetGroup to create location "/10_Legislation/Law" in targetDB
+            set theRecord to import POSIX path "{filepath}" to targetGroup
+            set tags of theRecord to {{"#주제/산업안전/법령", "$유형/법령", "{category}"}}
+            add custom meta data "law_monitor" for "sourceChannel" to theRecord
+            add custom meta data "external" for "dataOrigin" to theRecord
+            add custom meta data (current date) for "lastAIProcess" to theRecord
+        end if
+    end tell
+    '''
+    try:
+        run_applescript_inline(script)
+        logger.info(f"DEVONthink 임포트 완료: {law_name}")
+    except Exception as e:
+        logger.error(f"DEVONthink 임포트 실패 [{law_name}]: {e}")
+
+
+def run():
+    """메인 실행"""
+    logger.info("=== 법령 모니터링 시작 ===")
+
+    creds = load_credentials()
+    law_oc = creds.get("LAW_OC")
+    if not law_oc:
+        logger.error("LAW_OC 인증키가 설정되지 않았습니다. credentials.env를 확인하세요.")
+        sys.exit(1)
+
+    last_check = load_last_check()
+    changes_found = 0
+
+    for law in MONITORED_LAWS:
+        law_name = law["name"]
+        law_id = law["law_id"]
+        category = law["category"]
+
+        logger.info(f"확인 중: {law_name} ({law_id})")
+
+        info = fetch_law_info(law_oc, law_id)
+        if not info:
+            continue
+
+        # 시행일자 또는 공포일자로 변경 감지
+        announce_date = info.get("공포일자", info.get("시행일자", ""))
+        prev_date = last_check.get(law_id, "")
+
+        if announce_date and announce_date != prev_date:
+            logger.info(f"변경 감지: {law_name} — 공포일자 {announce_date} (이전: {prev_date or '없음'})")
+
+            # 법령 본문 다운로드
+            law_mst = info.get("법령MST", law_id)
+            text = fetch_law_text(law_oc, law_mst)
+            if text:
+                filepath = save_law_file(law_name, text)
+                import_to_devonthink(filepath, law_name, category)
+                changes_found += 1
+
+            last_check[law_id] = announce_date
+        else:
+            logger.debug(f"변경 없음: {law_name}")
+
+    save_last_check(last_check)
+    logger.info(f"=== 법령 모니터링 완료 — {changes_found}건 변경 감지 ===")
+
+
+if __name__ == "__main__":
+    run()

scripts/mailplus_archive.py

@@ -0,0 +1,209 @@
+#!/usr/bin/env python3
+"""
+MailPlus → DEVONthink Archive DB 이메일 수집
+- Synology MailPlus IMAP 접속
+- 마지막 동기화 이후 새 메일 가져오기
+- DEVONthink Archive DB 임포트
+"""
+
+import os
+import sys
+import imaplib
+import email
+from email.header import decode_header
+from datetime import datetime
+from pathlib import Path
+
+sys.path.insert(0, str(Path(__file__).parent))
+from pkm_utils import setup_logger, load_credentials, run_applescript_inline, DATA_DIR
+
+logger = setup_logger("mailplus")
+
+LAST_UID_FILE = DATA_DIR / "mailplus_last_uid.txt"
+MAIL_TMP_DIR = DATA_DIR / "mail_tmp"
+MAIL_TMP_DIR.mkdir(exist_ok=True)
+
+# 안전 관련 키워드 (dataOrigin 판별용)
+SAFETY_KEYWORDS = [
+    "안전", "위험", "사고", "재해", "점검", "보건", "화학물질",
+    "OSHA", "safety", "hazard", "incident", "KOSHA"
+]
+
+
+def decode_mime_header(value: str) -> str:
+    """MIME 헤더 디코딩"""
+    if not value:
+        return ""
+    decoded_parts = decode_header(value)
+    result = []
+    for part, charset in decoded_parts:
+        if isinstance(part, bytes):
+            result.append(part.decode(charset or "utf-8", errors="replace"))
+        else:
+            result.append(part)
+    return " ".join(result)
+
+
+def load_last_uid() -> int:
+    """마지막 처리 UID 로딩"""
+    if LAST_UID_FILE.exists():
+        return int(LAST_UID_FILE.read_text().strip())
+    return 0
+
+
+def save_last_uid(uid: int):
+    """마지막 처리 UID 저장"""
+    LAST_UID_FILE.write_text(str(uid))
+
+
+def detect_data_origin(subject: str, body: str) -> str:
+    """안전 키워드 감지로 dataOrigin 판별"""
+    text = (subject + " " + body).lower()
+    for kw in SAFETY_KEYWORDS:
+        if kw.lower() in text:
+            return "work"
+    return "external"
+
+
+def save_email_file(msg: email.message.Message, uid: int) -> Path:
+    """이메일을 .eml 파일로 저장"""
+    subject = decode_mime_header(msg.get("Subject", ""))
+    safe_subject = "".join(c if c.isalnum() or c in " _-" else "_" for c in subject)[:50]
+    date_str = datetime.now().strftime("%Y%m%d_%H%M%S")
+    filename = f"{date_str}_{uid}_{safe_subject}.eml"
+    filepath = MAIL_TMP_DIR / filename
+    with open(filepath, "wb") as f:
+        f.write(msg.as_bytes())
+    return filepath
+
+
+def get_email_body(msg: email.message.Message) -> str:
+    """이메일 본문 추출"""
+    body = ""
+    if msg.is_multipart():
+        for part in msg.walk():
+            if part.get_content_type() == "text/plain":
+                payload = part.get_payload(decode=True)
+                if payload:
+                    charset = part.get_content_charset() or "utf-8"
+                    body += payload.decode(charset, errors="replace")
+    else:
+        payload = msg.get_payload(decode=True)
+        if payload:
+            charset = msg.get_content_charset() or "utf-8"
+            body = payload.decode(charset, errors="replace")
+    return body[:2000]
+
+
+def import_to_devonthink(filepath: Path, subject: str, data_origin: str):
+    """DEVONthink Archive DB로 임포트"""
+    escaped_path = str(filepath).replace('"', '\\"')
+    escaped_subject = subject.replace('"', '\\"').replace("'", "\\'")
+    script = f'''
+    tell application id "DNtp"
+        set targetDB to missing value
+        repeat with db in databases
+            if name of db is "Archive" then
+                set targetDB to db
+                exit repeat
+            end if
+        end repeat
+
+        if targetDB is not missing value then
+            set targetGroup to create location "/Email" in targetDB
+            set theRecord to import POSIX path "{escaped_path}" to targetGroup
+            add custom meta data "email" for "sourceChannel" to theRecord
+            add custom meta data "{data_origin}" for "dataOrigin" to theRecord
+            add custom meta data (current date) for "lastAIProcess" to theRecord
+        end if
+    end tell
+    '''
+    try:
+        run_applescript_inline(script)
+        logger.info(f"DEVONthink 임포트: {subject[:40]}")
+    except Exception as e:
+        logger.error(f"DEVONthink 임포트 실패: {e}")
+
+
+def run():
+    """메인 실행"""
+    logger.info("=== MailPlus 이메일 수집 시작 ===")
+
+    creds = load_credentials()
+    host = creds.get("MAILPLUS_HOST")
+    port = int(creds.get("MAILPLUS_PORT", "993"))
+    user = creds.get("MAILPLUS_USER")
+    password = creds.get("MAILPLUS_PASS")
+
+    if not all([host, user, password]):
+        logger.error("MAILPLUS 접속 정보가 불완전합니다. credentials.env를 확인하세요.")
+        sys.exit(1)
+
+    last_uid = load_last_uid()
+    logger.info(f"마지막 처리 UID: {last_uid}")
+
+    try:
+        # IMAP SSL 접속
+        mail = imaplib.IMAP4_SSL(host, port)
+        mail.login(user, password)
+        mail.select("INBOX")
+        logger.info("IMAP 접속 성공")
+
+        # 마지막 UID 이후 메일 검색
+        if last_uid > 0:
+            status, data = mail.uid("search", None, f"UID {last_uid + 1}:*")
+        else:
+            # 최초 실행: 최근 7일치만
+            from datetime import timedelta
+            since = (datetime.now() - timedelta(days=7)).strftime("%d-%b-%Y")
+            status, data = mail.uid("search", None, f"SINCE {since}")
+
+        if status != "OK":
+            logger.error(f"메일 검색 실패: {status}")
+            mail.logout()
+            sys.exit(1)
+
+        uids = data[0].split()
+        logger.info(f"새 메일: {len(uids)}건")
+
+        max_uid = last_uid
+        imported = 0
+
+        for uid_bytes in uids:
+            uid = int(uid_bytes)
+            if uid <= last_uid:
+                continue
+
+            status, msg_data = mail.uid("fetch", uid_bytes, "(RFC822)")
+            if status != "OK":
+                continue
+
+            raw_email = msg_data[0][1]
+            msg = email.message_from_bytes(raw_email)
+
+            subject = decode_mime_header(msg.get("Subject", "(제목 없음)"))
+            body = get_email_body(msg)
+            data_origin = detect_data_origin(subject, body)
+
+            filepath = save_email_file(msg, uid)
+            import_to_devonthink(filepath, subject, data_origin)
+
+            max_uid = max(max_uid, uid)
+            imported += 1
+
+        if max_uid > last_uid:
+            save_last_uid(max_uid)
+
+        mail.logout()
+        logger.info(f"=== MailPlus 수집 완료 — {imported}건 임포트 ===")
+
+    except imaplib.IMAP4.error as e:
+        logger.error(f"IMAP 에러: {e}")
+        sys.exit(1)
+    except Exception as e:
+        logger.error(f"예상치 못한 에러: {e}", exc_info=True)
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    run()

scripts/pkm_daily_digest.py

@@ -0,0 +1,284 @@
+#!/usr/bin/env python3
+"""
+PKM 일일 다이제스트
+- DEVONthink 오늘 추가/수정 집계
+- law_monitor 법령 변경 건 파싱
+- OmniFocus 완료/추가/기한초과 집계
+- 상위 뉴스 Ollama 요약
+- OmniFocus 액션 자동 생성
+- 90일 지난 다이제스트 아카이브
+"""
+
+import os
+import sys
+import re
+from datetime import datetime, timedelta
+from pathlib import Path
+
+sys.path.insert(0, str(Path(__file__).parent))
+from pkm_utils import (
+    setup_logger, load_credentials, run_applescript_inline,
+    ollama_generate, count_log_errors, PROJECT_ROOT, LOGS_DIR, DATA_DIR
+)
+
+logger = setup_logger("digest")
+
+DIGEST_DIR = DATA_DIR / "digests"
+DIGEST_DIR.mkdir(exist_ok=True)
+
+
+def get_devonthink_stats() -> dict:
+    """DEVONthink 오늘 추가/수정 문서 집계"""
+    script = '''
+    tell application id "DNtp"
+        set today to current date
+        set time of today to 0
+        set stats to {}
+
+        repeat with db in databases
+            set dbName to name of db
+            set addedCount to count of (search "date:today" in db)
+            set modifiedCount to count of (search "modified:today" in db)
+
+            if addedCount > 0 or modifiedCount > 0 then
+                set end of stats to dbName & ":" & addedCount & ":" & modifiedCount
+            end if
+        end repeat
+
+        set AppleScript's text item delimiters to "|"
+        return stats as text
+    end tell
+    '''
+    try:
+        result = run_applescript_inline(script)
+        stats = {}
+        if result:
+            for item in result.split("|"):
+                parts = item.split(":")
+                if len(parts) == 3:
+                    stats[parts[0]] = {"added": int(parts[1]), "modified": int(parts[2])}
+        return stats
+    except Exception as e:
+        logger.error(f"DEVONthink 집계 실패: {e}")
+        return {}
+
+
+def get_omnifocus_stats() -> dict:
+    """OmniFocus 오늘 완료/추가/기한초과 집계"""
+    script = '''
+    tell application "OmniFocus"
+        tell default document
+            set today to current date
+            set time of today to 0
+            set tomorrow to today + 1 * days
+
+            set completedCount to count of (every flattened task whose completed is true and completion date ≥ today)
+            set addedCount to count of (every flattened task whose creation date ≥ today)
+            set overdueCount to count of (every flattened task whose completed is false and due date < today and due date is not missing value)
+
+            return (completedCount as text) & "|" & (addedCount as text) & "|" & (overdueCount as text)
+        end tell
+    end tell
+    '''
+    try:
+        result = run_applescript_inline(script)
+        parts = result.split("|")
+        return {
+            "completed": int(parts[0]) if len(parts) > 0 else 0,
+            "added": int(parts[1]) if len(parts) > 1 else 0,
+            "overdue": int(parts[2]) if len(parts) > 2 else 0,
+        }
+    except Exception as e:
+        logger.error(f"OmniFocus 집계 실패: {e}")
+        return {"completed": 0, "added": 0, "overdue": 0}
+
+
+def parse_law_changes() -> list:
+    """law_monitor 로그에서 오늘 법령 변경 건 파싱"""
+    log_file = LOGS_DIR / "law_monitor.log"
+    if not log_file.exists():
+        return []
+
+    today = datetime.now().strftime("%Y-%m-%d")
+    changes = []
+    with open(log_file, "r", encoding="utf-8") as f:
+        for line in f:
+            if today in line and "변경 감지" in line:
+                # "[2026-03-26 07:00:15] [law_monitor] [INFO] 변경 감지: 산업안전보건법 — 공포일자 ..."
+                match = re.search(r"변경 감지: (.+?)$", line)
+                if match:
+                    changes.append(match.group(1).strip())
+    return changes
+
+
+def get_inbox_count() -> int:
+    """DEVONthink Inbox 미처리 문서 수"""
+    script = '''
+    tell application id "DNtp"
+        repeat with db in databases
+            if name of db is "Inbox" then
+                return count of children of root group of db
+            end if
+        end repeat
+        return 0
+    end tell
+    '''
+    try:
+        return int(run_applescript_inline(script))
+    except:
+        return 0
+
+
+def create_omnifocus_task(task_name: str, note: str = "", flagged: bool = False):
+    """OmniFocus 작업 생성"""
+    flag_str = "true" if flagged else "false"
+    escaped_name = task_name.replace('"', '\\"')
+    escaped_note = note.replace('"', '\\"')
+    script = f'''
+    tell application "OmniFocus"
+        tell default document
+            make new inbox task with properties {{name:"{escaped_name}", note:"{escaped_note}", flagged:{flag_str}}}
+        end tell
+    end tell
+    '''
+    try:
+        run_applescript_inline(script)
+        logger.info(f"OmniFocus 작업 생성: {task_name}")
+    except Exception as e:
+        logger.error(f"OmniFocus 작업 생성 실패: {e}")
+
+
+def get_system_health() -> dict:
+    """각 모듈 로그의 최근 24시간 ERROR 카운트"""
+    modules = ["law_monitor", "mailplus", "digest", "embed", "auto_classify"]
+    health = {}
+    for mod in modules:
+        log_file = LOGS_DIR / f"{mod}.log"
+        health[mod] = count_log_errors(log_file, since_hours=24)
+    return health
+
+
+def generate_digest():
+    """다이제스트 생성"""
+    logger.info("=== Daily Digest 생성 시작 ===")
+    today = datetime.now()
+    date_str = today.strftime("%Y-%m-%d")
+
+    # 데이터 수집
+    dt_stats = get_devonthink_stats()
+    of_stats = get_omnifocus_stats()
+    law_changes = parse_law_changes()
+    inbox_count = get_inbox_count()
+    system_health = get_system_health()
+
+    # 마크다운 생성
+    md = f"# PKM Daily Digest — {date_str}\n\n"
+
+    # DEVONthink 현황
+    md += "## DEVONthink 변화\n\n"
+    if dt_stats:
+        md += "| DB | 신규 | 수정 |\n|---|---|---|\n"
+        total_added = 0
+        total_modified = 0
+        for db_name, counts in dt_stats.items():
+            md += f"| {db_name} | {counts['added']} | {counts['modified']} |\n"
+            total_added += counts["added"]
+            total_modified += counts["modified"]
+        md += f"| **합계** | **{total_added}** | **{total_modified}** |\n\n"
+    else:
+        md += "변화 없음\n\n"
+
+    # 법령 변경
+    md += "## 법령 변경\n\n"
+    if law_changes:
+        for change in law_changes:
+            md += f"- {change}\n"
+        md += "\n"
+    else:
+        md += "변경 없음\n\n"
+
+    # OmniFocus 현황
+    md += "## OmniFocus 현황\n\n"
+    md += f"- 완료: {of_stats['completed']}건\n"
+    md += f"- 신규: {of_stats['added']}건\n"
+    md += f"- 기한초과: {of_stats['overdue']}건\n\n"
+
+    # Inbox 상태
+    md += f"## Inbox 미처리: {inbox_count}건\n\n"
+
+    # 시스템 상태
+    md += "## 시스템 상태\n\n"
+    total_errors = sum(system_health.values())
+    if total_errors == 0:
+        md += "모든 모듈 정상\n\n"
+    else:
+        md += "| 모듈 | 에러 수 |\n|---|---|\n"
+        for mod, cnt in system_health.items():
+            status = f"**{cnt}**" if cnt > 0 else "0"
+            md += f"| {mod} | {status} |\n"
+        md += "\n"
+
+    # 파일 저장
+    digest_file = DIGEST_DIR / f"{date_str}_digest.md"
+    with open(digest_file, "w", encoding="utf-8") as f:
+        f.write(md)
+    logger.info(f"다이제스트 저장: {digest_file}")
+
+    # DEVONthink 저장
+    import_digest_to_devonthink(digest_file, date_str)
+
+    # OmniFocus 액션 자동 생성
+    if law_changes:
+        for change in law_changes:
+            create_omnifocus_task(f"법령 변경 검토: {change[:30]}", note=change)
+
+    if inbox_count >= 3:
+        create_omnifocus_task(f"Inbox 정리 ({inbox_count}건 미처리)", note="DEVONthink Inbox에 미분류 문서가 쌓여있습니다.")
+
+    if of_stats["overdue"] > 0:
+        create_omnifocus_task(f"기한초과 작업 처리 ({of_stats['overdue']}건)", flagged=True)
+
+    # 90일 지난 다이제스트 아카이브
+    archive_old_digests()
+
+    logger.info("=== Daily Digest 완료 ===")
+
+
+def import_digest_to_devonthink(filepath: Path, date_str: str):
+    """다이제스트를 DEVONthink에 저장"""
+    escaped_path = str(filepath).replace('"', '\\"')
+    script = f'''
+    tell application id "DNtp"
+        repeat with db in databases
+            if name of db is "00_Note_BOX" then
+                set targetGroup to create location "/Daily_Digest" in db
+                import POSIX path "{escaped_path}" to targetGroup
+                exit repeat
+            end if
+        end repeat
+    end tell
+    '''
+    try:
+        run_applescript_inline(script)
+    except Exception as e:
+        logger.error(f"DEVONthink 다이제스트 임포트 실패: {e}")
+
+
+def archive_old_digests():
+    """90일 지난 다이제스트 이동"""
+    cutoff = datetime.now() - timedelta(days=90)
+    for f in DIGEST_DIR.glob("*_digest.md"):
+        try:
+            date_part = f.stem.split("_digest")[0]
+            file_date = datetime.strptime(date_part, "%Y-%m-%d")
+            if file_date < cutoff:
+                archive_dir = DIGEST_DIR / "archive"
+                archive_dir.mkdir(exist_ok=True)
+                f.rename(archive_dir / f.name)
+                logger.info(f"아카이브: {f.name}")
+        except ValueError:
+            pass
+
+
+if __name__ == "__main__":
+    generate_digest()

scripts/pkm_utils.py

@@ -0,0 +1,138 @@
+"""
+PKM 시스템 공통 유틸리티
+- 로거 설정 (파일 + 콘솔)
+- credentials.env 로딩
+- osascript 호출 래퍼
+"""
+
+import os
+import sys
+import logging
+import subprocess
+from pathlib import Path
+from dotenv import load_dotenv
+
+# 프로젝트 루트 디렉토리
+PROJECT_ROOT = Path(__file__).parent.parent
+LOGS_DIR = PROJECT_ROOT / "logs"
+DATA_DIR = PROJECT_ROOT / "data"
+SCRIPTS_DIR = PROJECT_ROOT / "scripts"
+APPLESCRIPT_DIR = PROJECT_ROOT / "applescript"
+
+# 디렉토리 생성
+LOGS_DIR.mkdir(exist_ok=True)
+DATA_DIR.mkdir(exist_ok=True)
+
+
+def setup_logger(name: str) -> logging.Logger:
+    """모듈별 로거 설정 — 파일 + 콘솔 핸들러"""
+    logger = logging.getLogger(name)
+    if logger.handlers:
+        return logger  # 중복 핸들러 방지
+
+    logger.setLevel(logging.DEBUG)
+    fmt = logging.Formatter("[%(asctime)s] [%(name)s] [%(levelname)s] %(message)s",
+                            datefmt="%Y-%m-%d %H:%M:%S")
+
+    # 파일 핸들러
+    fh = logging.FileHandler(LOGS_DIR / f"{name}.log", encoding="utf-8")
+    fh.setLevel(logging.DEBUG)
+    fh.setFormatter(fmt)
+    logger.addHandler(fh)
+
+    # 콘솔 핸들러
+    ch = logging.StreamHandler(sys.stdout)
+    ch.setLevel(logging.INFO)
+    ch.setFormatter(fmt)
+    logger.addHandler(ch)
+
+    return logger
+
+
+def load_credentials() -> dict:
+    """~/.config/pkm/credentials.env 로딩 + 누락 키 경고"""
+    cred_path = Path.home() / ".config" / "pkm" / "credentials.env"
+    if not cred_path.exists():
+        # 폴백: 프로젝트 내 credentials.env (개발용)
+        cred_path = PROJECT_ROOT / "credentials.env"
+
+    if cred_path.exists():
+        load_dotenv(cred_path)
+    else:
+        print(f"[경고] credentials.env를 찾을 수 없습니다: {cred_path}")
+
+    keys = {
+        "CLAUDE_API_KEY": os.getenv("CLAUDE_API_KEY"),
+        "LAW_OC": os.getenv("LAW_OC"),
+        "NAS_DOMAIN": os.getenv("NAS_DOMAIN"),
+        "NAS_TAILSCALE_IP": os.getenv("NAS_TAILSCALE_IP"),
+        "NAS_PORT": os.getenv("NAS_PORT", "15001"),
+        "MAILPLUS_HOST": os.getenv("MAILPLUS_HOST"),
+        "MAILPLUS_PORT": os.getenv("MAILPLUS_PORT", "993"),
+        "MAILPLUS_USER": os.getenv("MAILPLUS_USER"),
+        "MAILPLUS_PASS": os.getenv("MAILPLUS_PASS"),
+        "GPU_SERVER_IP": os.getenv("GPU_SERVER_IP"),
+    }
+
+    missing = [k for k, v in keys.items() if not v and k not in ("GPU_SERVER_IP", "CLAUDE_API_KEY")]
+    if missing:
+        print(f"[경고] 누락된 인증 키: {', '.join(missing)}")
+
+    return keys
+
+
+def run_applescript(script_path: str, *args) -> str:
+    """osascript 호출 래퍼 + 에러 캡처"""
+    cmd = ["osascript", str(script_path)] + [str(a) for a in args]
+    try:
+        result = subprocess.run(cmd, capture_output=True, text=True, timeout=120)
+        if result.returncode != 0:
+            raise RuntimeError(f"AppleScript 에러: {result.stderr.strip()}")
+        return result.stdout.strip()
+    except subprocess.TimeoutExpired:
+        raise RuntimeError(f"AppleScript 타임아웃: {script_path}")
+
+
+def run_applescript_inline(script: str) -> str:
+    """인라인 AppleScript 실행"""
+    cmd = ["osascript", "-e", script]
+    try:
+        result = subprocess.run(cmd, capture_output=True, text=True, timeout=120)
+        if result.returncode != 0:
+            raise RuntimeError(f"AppleScript 에러: {result.stderr.strip()}")
+        return result.stdout.strip()
+    except subprocess.TimeoutExpired:
+        raise RuntimeError("AppleScript 타임아웃 (인라인)")
+
+
+def ollama_generate(prompt: str, model: str = "qwen3.5:35b-a3b-q4_K_M",
+                    host: str = "http://localhost:11434") -> str:
+    """Ollama API 호출"""
+    import requests
+    resp = requests.post(f"{host}/api/generate", json={
+        "model": model,
+        "prompt": prompt,
+        "stream": False
+    }, timeout=120)
+    resp.raise_for_status()
+    return resp.json().get("response", "")
+
+
+def count_log_errors(log_file: Path, since_hours: int = 24) -> int:
+    """로그 파일에서 최근 N시간 ERROR 카운트"""
+    from datetime import datetime, timedelta
+    if not log_file.exists():
+        return 0
+    cutoff = datetime.now() - timedelta(hours=since_hours)
+    count = 0
+    with open(log_file, "r", encoding="utf-8") as f:
+        for line in f:
+            if "[ERROR]" in line:
+                try:
+                    ts_str = line[1:20]  # [YYYY-MM-DD HH:MM:SS]
+                    ts = datetime.strptime(ts_str, "%Y-%m-%d %H:%M:%S")
+                    if ts >= cutoff:
+                        count += 1
+                except (ValueError, IndexError):
+                    count += 1
+    return count

scripts/prompts/classify_document.txt

@@ -0,0 +1,53 @@
+당신은 문서 분류 AI입니다. 아래 문서를 분석하고 반드시 JSON 형식으로만 응답하세요. 다른 텍스트는 출력하지 마세요.
+
+## 응답 형식
+{
+  "tags": ["태그1", "태그2", "태그3"],
+  "domain_db": "DB이름",
+  "sub_group": "하위그룹경로",
+  "sourceChannel": "유입경로",
+  "dataOrigin": "work 또는 external"
+}
+
+## 도메인 DB 선택지 (정확히 이 이름 사용)
+- 00_Note_BOX — 일반 메모, 스크랩, 잡다한 노트
+- 01_Philosophie — 철학, 사상, 인문학
+- 02_Language — 어학, 번역, 언어학
+- 03_Engineering — 공학 전반 기술 문서
+- 04_Industrial safety — 산업안전, 규정, 인증
+- 05_Programming — 개발, 코드, IT 기술
+- 07_General Book — 일반 도서, 독서 노트
+- 97_Production drawing — 생산 도면, CAD, 설계
+- 99_Reference Data — 범용 레퍼런스, 규격표
+- 99_Technicalkorea — 한국 기술 규정, 국내 기술 자료
+
+## 하위 그룹 경로 예시 (DB마다 다름)
+- 04_Industrial safety: 10_Legislation/Notice, 10_Legislation/Law, 20_Theory, 30_Papers, 40_Cases/Domestic, 50_Practice/Risk_Assessment, 50_Practice/Patrol_Inspection, 50_Practice/Education, 60_Compliance, 70_Safety_Manager, 80_Reference
+- 05_Programming: 10_Language, 20_Framework, 30_DevOps, 40_AI_ML
+- 03_Engineering: 10_Mechanical, 20_Electrical, 30_Network
+- 잘 모르겠으면: 00_Inbox
+
+## 태그 체계
+태그는 최대 5개, 한글 사용. 아래 계층 구조 중에서 선택:
+- @상태/: 처리중, 검토필요, 완료, 아카이브
+- #주제/기술/: 서버관리, 네트워크, AI-ML
+- #주제/산업안전/: 법령, 위험성평가, 순회점검, 안전교육, 사고사례, 신고보고, 안전관리자, 보건관리자
+- #주제/업무/: 프로젝트, 회의, 보고서
+- $유형/: 논문, 법령, 기사, 메모, 이메일, 채팅로그, 도면, 체크리스트
+- !우선순위/: 긴급, 중요, 참고
+
+## sourceChannel 값
+- tksafety: TKSafety API 업무 실적
+- devonagent: DEVONagent 자동 수집 뉴스
+- law_monitor: 법령 API 법령 변경
+- inbox_route: Inbox AI 분류 (이 프롬프트에 의한 분류)
+- email: MailPlus 이메일
+- web_clip: Web Clipper 스크랩
+- manual: 직접 추가
+
+## dataOrigin 값
+- work: 자사 업무 관련 (TK, 테크니컬코리아, 공장, 생산, 사내)
+- external: 외부 참고 자료 (뉴스, 논문, 법령, 일반 정보)
+
+## 분류 대상 문서
+{document_text}

tests/test_classify.py

@@ -0,0 +1,129 @@
+#!/usr/bin/env python3
+"""
+분류 프롬프트 단위 테스트
+5종 문서(법령, 뉴스, 논문, 메모, 이메일)로 분류 정확도 확인
+※ Ollama가 실행 중인 Mac mini에서 테스트해야 함
+"""
+
+import json
+import sys
+from pathlib import Path
+
+sys.path.insert(0, str(Path(__file__).parent.parent / "scripts"))
+from pkm_utils import ollama_generate, PROJECT_ROOT
+
+PROMPT_TEMPLATE = (PROJECT_ROOT / "scripts" / "prompts" / "classify_document.txt").read_text()
+
+# 테스트 문서 5종
+TEST_DOCUMENTS = [
+    {
+        "name": "법령 — 산업안전보건법 시행규칙 개정",
+        "text": "산업안전보건법 시행규칙 일부개정령안 입법예고. 고용노동부는 위험성평가에 관한 지침을 개정하여 사업장의 위험성평가 실시 주기를 연 1회에서 반기 1회로 강화하고, 위험성평가 결과의 기록 보존 기간을 3년에서 5년으로 확대하는 내용을 포함합니다.",
+        "expected_db": "04_Industrial safety",
+        "expected_origin": "external",
+    },
+    {
+        "name": "뉴스 — AI 기술 동향",
+        "text": "OpenAI가 GPT-5를 발표했습니다. 이번 모델은 멀티모달 기능이 대폭 강화되었으며, 코드 생성 능력이 기존 대비 40% 향상되었습니다. 특히 에이전트 기능이 추가되어 복잡한 작업을 자율적으로 수행할 수 있게 되었습니다.",
+        "expected_db": "05_Programming",
+        "expected_origin": "external",
+    },
+    {
+        "name": "논문 — 위험성평가 방법론",
+        "text": "A Literature Review on Risk Assessment Methodologies in Manufacturing Industry. This paper reviews the current state of risk assessment practices in Korean manufacturing facilities, comparing KOSHA guidelines with international standards including ISO 45001 and OSHA regulations.",
+        "expected_db": "04_Industrial safety",
+        "expected_origin": "external",
+    },
+    {
+        "name": "메모 — 업무 노트",
+        "text": "오늘 공장 순회점검에서 3층 용접 작업장의 환기 시설이 미작동 상태인 것을 확인했다. 시정조치 요청서를 작성하고 생산팀장에게 전달해야 한다. TODO: 시정조치 결과 확인 (3일 내)",
+        "expected_db": "04_Industrial safety",
+        "expected_origin": "work",
+    },
+    {
+        "name": "이메일 — 일반 업무",
+        "text": "안녕하세요 테크니컬코리아 안현기 과장님, 지난번 요청하신 용접기 부품 견적서를 첨부합니다. 납기는 발주 후 2주입니다. 감사합니다. - ○○산업 김철수 드림",
+        "expected_db": "99_Technicalkorea",
+        "expected_origin": "work",
+    },
+]
+
+
+def run_classify_test(doc: dict) -> dict:
+    """단일 문서 분류 테스트"""
+    prompt = PROMPT_TEMPLATE.replace("{document_text}", doc["text"])
+
+    try:
+        response = ollama_generate(prompt)
+        result = json.loads(response)
+
+        db_match = result.get("domain_db") == doc["expected_db"]
+        origin_match = result.get("dataOrigin") == doc["expected_origin"]
+
+        return {
+            "name": doc["name"],
+            "pass": db_match and origin_match,
+            "expected_db": doc["expected_db"],
+            "actual_db": result.get("domain_db"),
+            "expected_origin": doc["expected_origin"],
+            "actual_origin": result.get("dataOrigin"),
+            "tags": result.get("tags", []),
+            "sub_group": result.get("sub_group"),
+            "error": None,
+        }
+    except json.JSONDecodeError as e:
+        return {"name": doc["name"], "pass": False, "error": f"JSON 파싱 실패: {e}"}
+    except Exception as e:
+        return {"name": doc["name"], "pass": False, "error": str(e)}
+
+
+def main():
+    print("=" * 60)
+    print("PKM 문서 분류 테스트")
+    print("=" * 60)
+
+    results = []
+    for doc in TEST_DOCUMENTS:
+        print(f"\n테스트: {doc['name']}")
+        result = run_classify_test(doc)
+        results.append(result)
+
+        status = "PASS" if result["pass"] else "FAIL"
+        print(f"  [{status}]")
+        if result.get("error"):
+            print(f"  에러: {result['error']}")
+        else:
+            print(f"  DB: {result.get('actual_db')} (기대: {result.get('expected_db')})")
+            print(f"  Origin: {result.get('actual_origin')} (기대: {result.get('expected_origin')})")
+            print(f"  태그: {result.get('tags')}")
+            print(f"  그룹: {result.get('sub_group')}")
+
+    # 요약
+    passed = sum(1 for r in results if r["pass"])
+    total = len(results)
+    print(f"\n{'=' * 60}")
+    print(f"결과: {passed}/{total} 통과")
+    print("=" * 60)
+
+    # 리포트 저장
+    report_path = PROJECT_ROOT / "docs" / "test-report.md"
+    with open(report_path, "w", encoding="utf-8") as f:
+        f.write(f"# 분류 테스트 리포트\n\n")
+        f.write(f"실행일시: {__import__('datetime').datetime.now()}\n\n")
+        f.write(f"## 결과: {passed}/{total}\n\n")
+        for r in results:
+            status = "PASS" if r["pass"] else "FAIL"
+            f.write(f"### [{status}] {r['name']}\n")
+            if r.get("error"):
+                f.write(f"- 에러: {r['error']}\n")
+            else:
+                f.write(f"- DB: {r.get('actual_db')} (기대: {r.get('expected_db')})\n")
+                f.write(f"- Origin: {r.get('actual_origin')} (기대: {r.get('expected_origin')})\n")
+                f.write(f"- 태그: {r.get('tags')}\n\n")
+
+    print(f"\n리포트 저장: {report_path}")
+    sys.exit(0 if passed == total else 1)
+
+
+if __name__ == "__main__":
+    main()