merge: hygiene enforcement + full cleanup

2026-07-20 19:05:08 +08:00
parent 54c48dc5d7 4f83ee8a01
commit c545a82023
20 changed files with 1020 additions and 3 deletions
@@ -1,18 +1,23 @@
 # MoFin 开发规范

-> 版本: v1.0 | 更新: 2026-07-19 | 基于 AgentsMeeting 样板重构
+> 版本: v2.0 | 更新: 2026-07-20 | 基于 AgentsMeeting 样板重构 + 冗余事件复盘
 >
 > 📋 样板参考: [AgentsMeeting TEMPLATE-GUIDE.md](../AgentsMeeting/docs/TEMPLATE-GUIDE.md)

 ---

-## 五条红线
+## 十条红线

 1. **先读/写 Spec，再写代码** — 新增功能先写 spec 再实现；修改已有功能先读对应 spec 了解架构和约束再动手。没有 spec 的模块在 Dashboard 不可见，视为未完成
 2. **部署必验** — 部署后不打开 Dashboard F Tab 验证 = 部署未完成
 3. **不可见即不存在** — 组件不在 Dashboard 中显示 = 等于没部署。离线不告警 = 监控缺陷
 4. **实现后同步 Spec** — 每轮开发完毕后，必须将 `specs/{module}.json` 更新为与实际实现一致的状态。文档过期 = 等于没写
 5. **部署目标即验收标准** — 所有代码必须以部署目标环境（Linux 246）为基准编写和测试。禁止使用 Windows 专属 API（`tasklist`、`netstat`、`schtasks`、`wmic`）在 246 部署的代码中
+6. **单一事实源（SSOT）** — 每个文件全系统只有一个权威位置，其他位置只允许硬链接（同 inode），**禁止独立副本**。权威位置：`deploy/profile-scripts/`（cron 脚本）、`/home/hmo/MoFin/`（被 import 的库）、`deploy/bot/`（XMPP bot）。修改任何文件后若存在硬链接关系被破坏（scp/编辑器换 inode），必须立即跑 `deploy/profile-scripts/sync_profile_scripts.sh` 重建
+7. **数据路径必须绝对** — 引用数据文件/数据库时，必须写绝对路径并指向权威位置（`/home/hmo/MoFin/data/`）。**禁止**用 `Path(__file__).parent / "data"` 这类相对解析——同一个模块被硬链接到不同位置时会解析出不同的数据库（2026-07-20 三库事件的根因）
+8. **备份/遗留物禁止留在生产数据目录** — `.bak`、`decisions_backup_*`、迁移残留 JSON、废弃 DB，必须在迁移/变更完成时移到 `archive/`。生产数据目录（`MoFin/data` = `web-dashboard/data`）只放活文件。监控脚本扫描生产目录时，遗留物就是未来的假警报
+9. **死模块必须收尸** — 宣布模块废弃时，必须在同一轮操作中完成收尸六步：①杀进程 ②stop+disable systemd 服务 ③删 cron job ④归档脚本到 `archive/` ⑤归档数据文件 ⑥从期望矩阵/监控中移除。只说"已废弃"不收尸 = 没废弃（小果 bot 以 root 白跑 8 天 2.5GB 的教训）
+10. **监控查"活"不查"在"** — 健康检查必须验证**数据新鲜度**（DB 表 MAX(时间列)）而非"文件存在/进程存在"。文件 mtime、进程存活都不构成健康证据——数据 24h 不更新才是事故。禁止拿遗留文件的 mtime 当管道健康指标（"数据管道停滞14天"假警报的根因）

 ---

@@ -80,6 +85,7 @@ specs/{module}.json
 | dashboard | `specs/dashboard.json` | Dashboard 自身 | ✅ |
 | health | `specs/health.json` | 健康监控管线 | ✅ |
 | xmpp_monitor | `specs/xmpp_monitor.json` | XMPP 通信可观测性 | ✅ |
+| hygiene | `specs/hygiene.json` | 系统卫生审计（防冗余） | ✅ |
 | price_monitor | `specs/price_monitor.json` | 价格监控 cron | 📋 |
 | strategy_lifecycle | `specs/strategy_lifecycle.json` | 策略生命周期 | 📋 |

@@ -87,7 +93,36 @@ specs/{module}.json

 ---

-## 二、验证闭环
+## 二、文件位置宪法（2026-07-20 冗余事件后确立）
+
+| 内容类型 | 唯一权威位置 | 其他位置的合法形态 |
+|---------|-------------|------------------|
+| cron 脚本（被调度直接执行） | `deploy/profile-scripts/` | profile scripts 目录硬链接（经 `sync_profile_scripts.sh` 同步） |
+| 被 import 的库（mo_*/mofin_*/strategy_*/technical_*） | `/home/hmo/MoFin/`（根目录） | 禁止副本；deploy/profile-scripts 中的同名库文件只能是对 root 的硬链接 |
+| XMPP bot | `deploy/bot/` | `/home/hmo/xmpp_zhiwei_bot.py` 符号链接 |
+| Dashboard 服务 | `web-dashboard/server.py`（= `/home/hmo/MoFin/server.py` 硬链接） | — |
+| 数据文件/数据库 | `/home/hmo/MoFin/data/`（= `web-dashboard/data` 硬链接） | **禁止**任何第二个数据目录 |
+| 归档 | `archive/<主题>-<日期>/` | — |
+| 待销毁 | `trashbox/` | 定期人工清空 |
+
+**禁止出现的位置**：`MoFin/scripts/*.py`、`MoFin/` 根目录的 cron 脚本副本、`.hermes/*/scripts/data/`（任何 profile 本地 data 目录存业务数据）、`projects/` 下与生产同名的项目副本。
+
+### 死模块收尸清单（红线9 的执行版）
+
+```
+宣布模块 X 废弃时，同一轮操作内必须完成：
+□ 杀进程：pkill 或 systemctl stop（确认 ps 无残留）
+□ 服务：systemctl disable + rm unit 文件 + daemon-reload
+□ cron：两个 jobs.json 中删除 X 的 job，确认无残留
+□ 脚本：移到 archive/<模块>-retired-<日期>/
+□ 数据文件：同上
+□ 监控：从期望矩阵/健康检查注册表中移除 X
+□ 记录：CHANGELOG 写明收尸动作
+```
+
+---
+
+## 三、验证闭环

 ```
 ┌────────────┐     ┌──────────┐     ┌──────────┐
@@ -0,0 +1,45 @@
+import json, shutil, uuid
+from datetime import datetime
+
+jf = '/home/hmo/.hermes/profiles/position-analyst/cron/jobs.json'
+shutil.copy(jf, jf + '.bak-20260720-hygiene')
+d = json.load(open(jf))
+is_list = isinstance(d, list)
+jobs = d if is_list else d.get('jobs', [])
+
+# 防重复
+if any(j.get('script') == 'system_hygiene_audit.py' for j in jobs):
+    print('job already exists')
+else:
+    job = {
+        "id": uuid.uuid4().hex[:12],
+        "name": "系统卫生审计-每周",
+        "prompt": "",
+        "skills": [],
+        "skill": None,
+        "model": None,
+        "provider": None,
+        "base_url": None,
+        "script": "system_hygiene_audit.py",
+        "no_agent": True,
+        "context_from": None,
+        "schedule": {"kind": "cron", "expr": "30 7 * * 1", "display": "30 7 * * 1"},
+        "schedule_display": "30 7 * * 1",
+        "repeat": {"times": None, "completed": 0},
+        "enabled": True,
+        "state": "scheduled",
+        "paused_at": None,
+        "paused_reason": None,
+        "created_at": datetime.now().isoformat(),
+        "next_run_at": "2026-07-21T07:30:00+08:00",
+        "last_run_at": None,
+        "last_status": None,
+    }
+    jobs.append(job)
+    if is_list:
+        json.dump(jobs, open(jf, 'w'), ensure_ascii=False, indent=2)
+    else:
+        d['jobs'] = jobs
+        json.dump(d, open(jf, 'w'), ensure_ascii=False, indent=2)
+    print('added job: 系统卫生审计-每周 (Mon 07:30)')
+print('total jobs:', len(jobs))
@@ -0,0 +1,58 @@
+import os, shutil, glob
+from datetime import datetime
+
+ARCHIVE = f'/home/hmo/MoFin/archive/json-legacy-20260720'
+os.makedirs(ARCHIVE, exist_ok=True)
+
+# 保留的活文件（在用，不动）
+KEEP = {
+    'evaluation.json',           # evaluator 在读
+    'accuracy_stats.json',       # strategy_feedback 在读
+    'format_error_library.json', # 在用 (9d)
+    'candidate_pool.json',       # 在用 (3d)
+    'pipeline_registry.json',    # 在用 (3d)
+    # 今天还在写的
+    'growth_registry.json', 'hardcode_audit.json', 'health_checklist.json',
+    'macro_divergence_state.json', 'macro_risk_state.json', 'market.json',
+    'mofin_health.json', 'portfolio.json', 'preflight_result.json',
+    'price_history.json', 'scanner_state.json', 'state.db',
+    'strategy_staleness_report.json', 'system_audit_report.json',
+    'mofin.db', 'mofin.db-shm', 'mofin.db-wal',
+    'analyst-knowledge-log.md', 'mofin_health.html', 'evaluation_input.json',
+    'push_cooldown.json', 'system_audit.json', 'system_inventory.json',
+    'watchlist.json.bak2',
+}
+
+DATA = '/home/hmo/web-dashboard/data'  # = MoFin/data 硬链接
+moved = 0
+for f in sorted(os.listdir(DATA)):
+    p = os.path.join(DATA, f)
+    if not os.path.isfile(p):
+        continue
+    if f in KEEP:
+        continue
+    if f.startswith('.'):
+        continue
+    # 只归档 json/db/txt/log/md 类数据文件，别的不动
+    if not any(f.endswith(ext) for ext in ('.json', '.db', '.txt')):
+        continue
+    # 归档目标明确为遗留：>7天未修改
+    age_d = (datetime.now().timestamp() - os.path.getmtime(p)) / 86400
+    if age_d < 7:
+        continue
+    shutil.move(p, os.path.join(ARCHIVE, f))
+    moved += 1
+    print(f'  archived: {f} ({age_d:.0f}d)')
+
+print(f'\narchived {moved} files -> {ARCHIVE}')
+
+# 废弃小库
+print('\n=== 废弃小库 ===')
+for f in ['market.db', 'market_data.db', 'stock_analysis.db']:
+    p = os.path.join(DATA, f)
+    if os.path.exists(p):
+        print(f'  {f}: 已在归档中' if not os.path.exists(p) else '')
+# 上面已经按 >7d 规则归档了（27d/21d 的 db 文件），确认：
+for f in ['market.db', 'market_data.db', 'stock_analysis.db']:
+    archived = os.path.join(ARCHIVE, f)
+    print(f'  {f}: {"ARCHIVED" if os.path.exists(archived) else "still present!"}')
@@ -0,0 +1,30 @@
+import os, shutil, glob
+from datetime import datetime
+
+ARCHIVE = '/home/hmo/MoFin/archive/old-files-20260720'
+os.makedirs(ARCHIVE, exist_ok=True)
+
+# 1. /home/hmo 根目录旧日志（>10天）
+count = 0
+for f in glob.glob('/home/hmo/*.log'):
+    age_d = (datetime.now().timestamp() - os.path.getmtime(f)) / 86400
+    if age_d > 10:
+        shutil.move(f, os.path.join(ARCHIVE, os.path.basename(f)))
+        count += 1
+        print(f'  log: {os.path.basename(f)} ({age_d:.0f}d)')
+print(f'logs archived: {count}')
+
+# 2. 旧 bot 文件
+for f in ['/home/hmo/xmpp_zhiwei_bot.py.bak', '/home/hmo/run_zhiwei_bot.py', '/home/hmo/xmpp_bot_rest.py']:
+    if os.path.exists(f):
+        shutil.move(f, os.path.join(ARCHIVE, os.path.basename(f)))
+        print(f'  bot: {os.path.basename(f)}')
+
+# 3. 数据目录 bak 文件
+for f in ['/home/hmo/web-dashboard/data/watchlist.json.bak2',
+          '/home/hmo/.hermes/profiles/position-analyst/data/mofin.db.bak']:
+    if os.path.exists(f):
+        shutil.move(f, os.path.join(ARCHIVE, os.path.basename(f)))
+        print(f'  bak: {os.path.basename(f)}')
+
+print('DONE ->', ARCHIVE)
@@ -0,0 +1,84 @@
+#!/usr/bin/env python3
+"""audit_data.py — 数据存储碎片化审计：三个数据目录 + 遗留JSON + 旧库"""
+import os, json, sqlite3
+from datetime import datetime
+
+DATA_DIRS = [
+    '/home/hmo/MoFin/data',
+    '/home/hmo/web-dashboard/data',
+    '/home/hmo/.hermes/profiles/position-analyst/scripts/data',
+    '/home/hmo/.hermes/profiles/position-analyst/data',
+    '/home/hmo/.hermes/data',
+]
+
+now = datetime.now()
+
+print("=== 1. 各数据目录内容与时效 ===")
+for dd in DATA_DIRS:
+    if not os.path.isdir(dd):
+        print(f"\n{dd}: 不存在")
+        continue
+    print(f"\n{dd}:")
+    try:
+        entries = []
+        for f in sorted(os.listdir(dd)):
+            p = os.path.join(dd, f)
+            if os.path.isfile(p):
+                mt = datetime.fromtimestamp(os.path.getmtime(p))
+                age_h = (now - mt).total_seconds() / 3600
+                entries.append((f, os.path.getsize(p), mt, age_h))
+        # 只显示 >7天 或 >100KB 的
+        for f, size, mt, age_h in entries:
+            flag = '🔴' if age_h > 24*14 else ('🟡' if age_h > 24*3 else '🟢')
+            print(f"  {flag} {f:45s} {size//1024:6d}KB  {mt.strftime('%m-%d %H:%M')} ({age_h/24:.0f}d)")
+    except Exception as e:
+        print(f"  ERR: {e}")
+
+print("\n=== 2. 三个 mofin.db 对比 ===")
+for label, path in [('canonical', '/home/hmo/MoFin/data/mofin.db'),
+                    ('profile-local', '/home/hmo/.hermes/profiles/position-analyst/scripts/data/mofin.db'),
+                    ('profile-data', '/home/hmo/.hermes/profiles/position-analyst/data/mofin.db')]:
+    if not os.path.exists(path):
+        print(f"  {label}: 不存在")
+        continue
+    c = sqlite3.connect(path, timeout=5)
+    tables = {r[0]: r[1] for r in c.execute(
+        "SELECT name, (SELECT COUNT(*) FROM sqlite_master m2 WHERE m2.name=m1.name) FROM sqlite_master m1 WHERE type='table'")}
+    cnts = {}
+    for t in tables:
+        try:
+            cnts[t] = c.execute(f"SELECT COUNT(*) FROM {t}").fetchone()[0]
+        except Exception:
+            cnts[t] = -1
+    nonempty = {k: v for k, v in cnts.items() if v > 0}
+    print(f"  {label} ({os.path.getsize(path)//1024}KB): {len(nonempty)} 张非空表")
+    for t, n in sorted(nonempty.items(), key=lambda x: -x[1])[:8]:
+        print(f"    {t}: {n}")
+    c.close()
+
+print("\n=== 3. 遗留 JSON 文件（web-dashboard/data，按最后修改排序）===")
+wd = '/home/hmo/web-dashboard/data'
+jfiles = []
+for f in os.listdir(wd):
+    if f.endswith('.json') and os.path.isfile(os.path.join(wd, f)):
+        p = os.path.join(wd, f)
+        mt = datetime.fromtimestamp(os.path.getmtime(p))
+        jfiles.append((f, os.path.getsize(p), mt))
+jfiles.sort(key=lambda x: x[2])
+for f, size, mt in jfiles:
+    age_d = (now - mt).total_seconds() / 86400
+    flag = '🔴' if age_d > 14 else ('🟡' if age_d > 3 else '🟢')
+    print(f"  {flag} {f:45s} {size//1024:5d}KB  {mt.strftime('%m-%d')} ({age_d:.0f}d前)")
+
+print("\n=== 4. trashbox / archive / 旧日志 ===")
+for d in ['/home/hmo/trashbox', '/home/hmo/MoFin/archive', '/home/hmo/MoFin/data/archive']:
+    if os.path.isdir(d):
+        total = sum(os.path.getsize(os.path.join(r, f)) for r, _, fs in os.walk(d) for f in fs if os.path.isfile(os.path.join(r, f)))
+        cnt = sum(len(fs) for _, _, fs in os.walk(d))
+        print(f"  {d}: {cnt} 个文件, {total//1024//1024}MB")
+logs = [f for f in os.listdir('/home/hmo') if f.endswith('.log') and os.path.isfile(f'/home/hmo/{f}')]
+print(f"\n  /home/hmo 根目录 .log 文件: {len(logs)} 个")
+for f in sorted(logs):
+    p = f'/home/hmo/{f}'
+    mt = datetime.fromtimestamp(os.path.getmtime(p))
+    print(f"    {f} {os.path.getsize(p)//1024}KB {mt.strftime('%m-%d')}")
@@ -0,0 +1,92 @@
+#!/usr/bin/env python3
+"""audit_deadcode.py — 废弃模块/死代码审计：小果生态、旧bot、疑似废弃脚本"""
+import os, subprocess, json
+from datetime import datetime
+
+def sh(cmd, timeout=10):
+    try:
+        r = subprocess.run(cmd, shell=True, capture_output=True, text=True, timeout=timeout)
+        return r.stdout.strip()
+    except Exception as e:
+        return f"ERR: {e}"
+
+print("=== 1. 小果(xiaoguo)生态残留 ===")
+print("--- 运行中的 xiaoguo 进程 ---")
+print(sh("ps aux | grep -i xiaoguo | grep -v grep"))
+print("\n--- xiaoguo 相关文件 ---")
+print(sh("ls -la /home/hmo/MoFin/deploy/profile-scripts/ 2>/dev/null | grep -i xiaoguo; ls /home/hmo/MoFin/ | grep -i xiaoguo; ls /home/hmo/MoFin/scripts/ 2>/dev/null | grep -i xiaoguo; ls /home/hmo/AgentsMeeting/ 2>/dev/null | grep -i xiaoguo | head -5"))
+print("\n--- xiaoguo 相关数据文件 ---")
+print(sh("ls -la /home/hmo/web-dashboard/data/ | grep -i xiaoguo; ls -la /home/hmo/MoFin/data/ 2>/dev/null | grep -i xiaoguo"))
+print("\n--- xiaoguo 相关 cron ---")
+for jf, prof in [('/home/hmo/.hermes/profiles/position-analyst/cron/jobs.json', 'pa'), ('/home/hmo/.hermes/cron/jobs.json', 'default')]:
+    try:
+        d = json.load(open(jf))
+        jobs = d if isinstance(d, list) else d.get('jobs', [])
+        for j in jobs:
+            if 'xiaoguo' in str(j.get('script', '')).lower() or '小果' in str(j.get('name', '')):
+                print(f"  [{prof}] {j.get('name')} script={j.get('script')} enabled={j.get('enabled')} last={str(j.get('last_run_at'))[:10]} status={j.get('last_status')}")
+    except Exception as e:
+        print(f"  {prof}: {e}")
+
+print("\n=== 2. 旧 bot 文件 ===")
+for f in ['/home/hmo/xmpp_zhiwei_bot.py.bak', '/home/hmo/run_zhiwei_bot.py', '/home/hmo/xmpp_bot_rest.py',
+          '/home/hmo/xmpp_xiaoguo_bot.py', '/home/hmo/xmpp_mohe_bot.py', '/home/hmo/xmpp_zhiwei_bot.py']:
+    if os.path.exists(f):
+        p = f
+        mt = datetime.fromtimestamp(os.path.getmtime(p))
+        link = os.path.islink(p)
+        tgt = os.readlink(p) if link else ''
+        print(f"  {f} {'symlink->'+tgt if link else str(os.path.getsize(p)//1024)+'KB'} {mt.strftime('%m-%d')}")
+
+print("\n=== 3. AgentsMeeting 目录与 MoFin 的关系（是否重复项目）===")
+print(sh("du -sh /home/hmo/AgentsMeeting /home/hmo/MoFin /home/hmo/web-dashboard /home/hmo/projects/AgentsMeeting /home/hmo/projects/MoFin 2>/dev/null"))
+print("\n--- AgentsMeeting 里和 MoFin 同名的脚本 ---")
+am = '/home/hmo/AgentsMeeting'
+if os.path.isdir(am):
+    am_scripts = set()
+    for r, _, fs in os.walk(am):
+        if 'venv' in r or 'node_modules' in r or '.git' in r:
+            continue
+        for f in fs:
+            if f.endswith('.py'):
+                am_scripts.add(f)
+    mofin_scripts = set()
+    for d in ['/home/hmo/MoFin', '/home/hmo/MoFin/scripts', '/home/hmo/MoFin/deploy/profile-scripts']:
+        if os.path.isdir(d):
+            mofin_scripts.update(f for f in os.listdir(d) if f.endswith('.py'))
+    common = sorted(am_scripts & mofin_scripts)
+    print(f"  同名 .py 文件: {len(common)} 个")
+    for f in common[:20]:
+        print(f"    {f}")
+
+print("\n=== 4. 超过30天未修改的可疑废弃脚本（deploy/profile-scripts 中）===")
+now = datetime.now()
+dd = '/home/hmo/MoFin/deploy/profile-scripts'
+for f in sorted(os.listdir(dd)):
+    if not f.endswith('.py'):
+        continue
+    p = os.path.join(dd, f)
+    mt = datetime.fromtimestamp(os.path.getmtime(p))
+    age = (now - mt).total_seconds() / 86400
+    if age > 30:
+        print(f"  {f:45s} {mt.strftime('%m-%d')} ({age:.0f}d)")
+
+print("\n=== 5. scripts/ 下的工具脚本（非 cron 调度，疑似一次性）===")
+# 哪些脚本不在 cron jobs 里
+cron_scripts = set()
+for jf in ['/home/hmo/.hermes/profiles/position-analyst/cron/jobs.json', '/home/hmo/.hermes/cron/jobs.json']:
+    try:
+        d = json.load(open(jf))
+        jobs = d if isinstance(d, list) else d.get('jobs', [])
+        for j in jobs:
+            if j.get('script'):
+                cron_scripts.add(j['script'])
+    except Exception:
+        pass
+for d in ['/home/hmo/MoFin/deploy/profile-scripts']:
+    for f in sorted(os.listdir(d)):
+        if f.endswith('.py') and f not in cron_scripts:
+            p = os.path.join(d, f)
+            mt = datetime.fromtimestamp(os.path.getmtime(p))
+            age = (now - mt).total_seconds() / 86400
+            print(f"  {f:45s} 不在cron中 ({age:.0f}d)")
@@ -0,0 +1,84 @@
+#!/usr/bin/env python3
+"""audit_duplication.py — 扫描 MoFin 相关目录的文件重复情况"""
+import os, hashlib, json
+from collections import defaultdict
+
+LOCATIONS = [
+    '/home/hmo/MoFin',
+    '/home/hmo/MoFin/scripts',
+    '/home/hmo/MoFin/deploy/profile-scripts',
+    '/home/hmo/MoFin/deploy/bot',
+    '/home/hmo/.hermes/profiles/position-analyst/scripts',
+    '/home/hmo/.hermes/scripts',
+    '/home/hmo/web-dashboard',
+]
+
+def md5(p):
+    try:
+        with open(p, 'rb') as f:
+            return hashlib.md5(f.read()).hexdigest()[:10]
+    except Exception:
+        return 'ERR'
+
+def ino(p):
+    try:
+        return os.stat(p).st_ino
+    except Exception:
+        return 0
+
+# 收集所有 .py 文件
+files = defaultdict(list)  # name -> [(loc, path, md5, inode, is_link)]
+for loc in LOCATIONS:
+    if not os.path.isdir(loc):
+        continue
+    for f in os.listdir(loc):
+        if not f.endswith('.py'):
+            continue
+        p = os.path.join(loc, f)
+        if not os.path.isfile(p):
+            continue
+        files[f].append({
+            'loc': loc, 'path': p, 'md5': md5(p), 'inode': ino(p),
+            'is_link': os.path.islink(p),
+            'size': os.path.getsize(p),
+            'mtime': int(os.path.getmtime(p)),
+        })
+
+print("=== 重复文件（同名出现在2+位置）===")
+dups = {k: v for k, v in files.items() if len(v) > 1}
+identical = 0
+hardlinked = 0
+diverged = 0
+for name in sorted(dups):
+    entries = dups[name]
+    md5s = set(e['md5'] for e in entries)
+    inodes = set(e['inode'] for e in entries)
+    if len(inodes) == 1:
+        status = 'HARDLINK(同一文件)'
+        hardlinked += 1
+    elif len(md5s) == 1:
+        status = 'COPY(内容相同,多份独立)'
+        identical += 1
+    else:
+        status = 'DIVERGED(内容不同!)'
+        diverged += 1
+    print(f"{status} {name}")
+    for e in entries:
+        print(f"    {e['path']} md5={e['md5']} ino={e['inode']} size={e['size']}")
+
+print(f"\n汇总: {len(dups)} 个重复文件名 | hardlink={hardlinked} 内容相同副本={identical} 内容分叉={diverged}")
+
+print("\n=== deploy/profile-scripts 中有但 position-analyst/scripts 中缺失的 ===")
+pa_dir = '/home/hmo/.hermes/profiles/position-analyst/scripts'
+deploy_dir = '/home/hmo/MoFin/deploy/profile-scripts'
+pa_files = set(os.listdir(pa_dir)) if os.path.isdir(pa_dir) else set()
+for f in sorted(os.listdir(deploy_dir)):
+    if f.endswith('.py') and f not in pa_files:
+        print(f"  {f}")
+
+print("\n=== position-analyst/scripts 中有但 deploy 中没有的（可能孤儿）===")
+deploy_files = set(os.listdir(deploy_dir))
+for f in sorted(pa_files):
+    if f.endswith('.py') and f not in deploy_files:
+        p = os.path.join(pa_dir, f)
+        print(f"  {f} size={os.path.getsize(p)}")
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+"""audit_runtime.py — 运行中的进程/服务/cron 审计"""
+import subprocess, json, os
+
+def sh(cmd, timeout=10):
+    try:
+        r = subprocess.run(cmd, shell=True, capture_output=True, text=True, timeout=timeout)
+        return r.stdout.strip()
+    except Exception as e:
+        return f"ERR: {e}"
+
+print("=== 1. 长期运行的 Python 进程（>1天）===")
+out = sh("ps -eo pid,etime,user,args --sort=-etime | grep -E 'python|node' | grep -v grep | head -25")
+print(out)
+
+print("\n=== 2. systemd 服务状态（全部 mofin/hermes/xmpp/ejabberd 相关）===")
+out = sh("systemctl list-units --all --type=service 2>/dev/null | grep -iE 'mofin|hermes|xmpp|ejabberd|wechat|gitea|kanban|wiki|stock|xiaoguo|zhiwei|mohe' | head -25")
+print(out)
+print("\n--- user services ---")
+out = sh("systemctl --user list-units --all --type=service 2>/dev/null | grep -iE 'hermes|xmpp|mofin|wiki' | head -10")
+print(out)
+
+print("\n=== 3. DISABLED cron jobs（hermes 两个 profile）===")
+for jf, prof in [('/home/hmo/.hermes/profiles/position-analyst/cron/jobs.json', 'position-analyst'),
+                 ('/home/hmo/.hermes/cron/jobs.json', 'default')]:
+    try:
+        d = json.load(open(jf))
+        jobs = d if isinstance(d, list) else d.get('jobs', [])
+        for j in jobs:
+            if not j.get('enabled', True):
+                lr = str(j.get('last_run_at') or '?')[:10]
+                print(f"  [{prof}] {j.get('name')} | script={j.get('script')} | last={lr} | paused_reason={j.get('paused_reason')}")
+    except Exception as e:
+        print(f"  {prof}: {e}")
+
+print("\n=== 4. 系统 crontab 全文（非注释行）===")
+out = sh("crontab -l | grep -vE '^\\s*#' | grep -vE '^\\s*$'")
+print(out)
+
+print("\n=== 5. 僵尸/异常进程（root跑的python、很老的进程）===")
+out = sh("ps -eo pid,etime,user,args | grep -E '^\\s*\\S+\\s+\\S+\\s+root' | grep python | grep -v grep")
+print(out if out else "(none)")
+
+print("\n=== 6. 监听端口清单 ===")
+out = sh("ss -tlnp 2>/dev/null | grep -E '8642|8643|8645|8646|8899|5801|5802|5803|5804|5805|5807|5808|5810|9090|9580|5222|5443|5280|19088|19099|9877|9878' | awk '{print $4, $6}'")
+print(out)
@@ -0,0 +1,47 @@
+import subprocess, os, re
+from datetime import datetime
+
+DIVERGED = """advice_reconciliation.py branch_scanner.py bulk_strategy_regenerate.py
+collect_evaluation_data.py cron_to_xmpp.py market_insight.py market_screener.py market_watch.py
+memory_guardian.py mo_provider.py mofin_collect.py mofin_health.py mofin_news.py multi_timeframe.py
+premarket_full_review.py promote_candidates.py prune_branches.py server.py stock_profile.py
+stock_sector_enrich.py strategy_evaluator.py strategy_feedback.py strategy_lifecycle.py strategy_tree.py
+system_audit.py system_health_check.py technical_analysis.py trend_detector.py
+xiaoguo_news_processor.py xiaoguo_scanner.py xmpp_agent_core.py""".split()
+
+SEARCH_DIRS = ['/home/hmo/MoFin/deploy/profile-scripts', '/home/hmo/.hermes/profiles/position-analyst/scripts', '/home/hmo/.hermes/scripts']
+
+now = datetime.now()
+
+print(f"{'file':35s} {'imported_by':40s} {'cron?':5s} {'root_mtime':12s} {'deploy_mtime':12s} verdict")
+print('-' * 130)
+
+for f in DIVERGED:
+    stem = f[:-3]
+    # 谁 import 它
+    r = subprocess.run(
+        f"grep -rln -E '(^|\\s)(from|import)\\s+{stem}(\\s|$|\\.)' /home/hmo/MoFin/deploy/profile-scripts /home/hmo/.hermes/profiles/position-analyst/scripts /home/hmo/MoFin --include='*.py' 2>/dev/null | grep -v venv | grep -v '{f}' | head -3",
+        shell=True, capture_output=True, text=True, timeout=15)
+    importers = [os.path.basename(x) for x in r.stdout.splitlines() if x.strip() and f not in x]
+    imported = ','.join(importers[:3]) if importers else '-'
+
+    # 是否在 cron 里被直接执行
+    in_cron = subprocess.run(
+        f"grep -l '\"script\": \"{f}\"' /home/hmo/.hermes/profiles/position-analyst/cron/jobs.json /home/hmo/.hermes/cron/jobs.json 2>/dev/null",
+        shell=True, capture_output=True, text=True, timeout=5)
+    is_cron = 'Y' if in_cron.stdout.strip() else '-'
+
+    # mtimes
+    root_p = f'/home/hmo/MoFin/{f}'
+    deploy_p = f'/home/hmo/MoFin/deploy/profile-scripts/{f}'
+    rm = datetime.fromtimestamp(os.path.getmtime(root_p)).strftime('%m-%d') if os.path.exists(root_p) else '-'
+    dm = datetime.fromtimestamp(os.path.getmtime(deploy_p)).strftime('%m-%d') if os.path.exists(deploy_p) else '-'
+
+    # verdict
+    if importers:
+        verdict = 'LIBRARY->root为准'
+    elif is_cron == 'Y':
+        verdict = 'CRON->deploy为准'
+    else:
+        verdict = '待查'
+    print(f"{f:35s} {imported:40s} {is_cron:5s} {rm:12s} {dm:12s} {verdict}")
@@ -0,0 +1,30 @@
+import json, shutil
+
+REMOVE = {
+    'pa': ['MoFin 盘前中监控', 'MoFin 午后监控', '自选买入区提醒', '候选股自动推广-盘中'],
+    'default': ['持仓情报-盘中', '快速盯盘-15分钟', '价格监控-1分钟', 'PM-项目跟进'],
+}
+FILES = {
+    'pa': '/home/hmo/.hermes/profiles/position-analyst/cron/jobs.json',
+    'default': '/home/hmo/.hermes/cron/jobs.json',
+}
+
+for prof, jf in FILES.items():
+    shutil.copy(jf, jf + '.bak-20260720-disabled-cleanup')
+    d = json.load(open(jf))
+    is_list = isinstance(d, list)
+    jobs = d if is_list else d.get('jobs', [])
+    removed = [j.get('name') for j in jobs if j.get('name') in REMOVE[prof]]
+    kept = [j for j in jobs if j.get('name') not in REMOVE[prof]]
+    if removed:
+        if is_list:
+            json.dump(kept, open(jf, 'w'), ensure_ascii=False, indent=2)
+        else:
+            d['jobs'] = kept
+            json.dump(d, open(jf, 'w'), ensure_ascii=False, indent=2)
+    print(f'{prof}: removed {removed}')
+    # 剩余禁用任务
+    still_disabled = [j.get('name') for j in kept if not j.get('enabled', True)]
+    print(f'  remaining disabled: {still_disabled}')
+    total = len(kept)
+    print(f'  total jobs now: {total}')
@@ -0,0 +1,27 @@
+import subprocess
+
+# 更新 crontab：cron_to_xmpp.py 改指 deploy 路径（web-dashboard 副本已归档）
+r = subprocess.run(['crontab', '-l'], capture_output=True, text=True)
+lines = r.stdout.splitlines()
+new_lines = []
+for line in lines:
+    if 'cron_to_xmpp.py' in line and 'web-dashboard' in line:
+        new_line = line.replace(
+            'cd /home/hmo/web-dashboard && python3 cron_to_xmpp.py',
+            'cd /home/hmo/MoFin/deploy/profile-scripts && python3 cron_to_xmpp.py')
+        new_lines.append(new_line)
+        print('CHANGED:')
+        print('  old:', line)
+        print('  new:', new_line)
+    else:
+        new_lines.append(line)
+
+p = subprocess.run(['crontab', '-'], input='\n'.join(new_lines) + '\n',
+                   capture_output=True, text=True)
+print('crontab updated, rc=', p.returncode)
+
+# 验证 deploy 版本能跑（dry check: import + 语法）
+r2 = subprocess.run(
+    "cd /home/hmo/MoFin/deploy/profile-scripts && python3 -c 'import ast; ast.parse(open(\"cron_to_xmpp.py\").read()); print(\"syntax OK\")'",
+    shell=True, capture_output=True, text=True, timeout=10)
+print(r2.stdout.strip() or r2.stderr.strip()[:200])
@@ -0,0 +1,17 @@
+import subprocess
+
+r = subprocess.run(['crontab', '-l'], capture_output=True, text=True)
+lines = r.stdout.splitlines()
+new_lines = []
+for line in lines:
+    nl = line
+    if 'cd /home/hmo/MoFin &&' in line and ('market_watch.py' in line or 'market_screener.py' in line):
+        nl = line.replace('cd /home/hmo/MoFin &&', 'cd /home/hmo/MoFin/deploy/profile-scripts &&')
+        print('CHANGED:')
+        print('  old:', line)
+        print('  new:', nl)
+    new_lines.append(nl)
+
+p = subprocess.run(['crontab', '-'], input='\n'.join(new_lines) + '\n',
+                   capture_output=True, text=True)
+print('rc=', p.returncode)
@@ -0,0 +1,33 @@
+import sqlite3
+from datetime import datetime
+
+THIRD = '/home/hmo/.hermes/profiles/position-analyst/scripts/data/mofin.db'
+MAIN = '/home/hmo/MoFin/data/mofin.db'
+
+t = sqlite3.connect(THIRD, timeout=10)
+m = sqlite3.connect(MAIN, timeout=30)
+m.execute('PRAGMA busy_timeout=30000')
+
+for table in ['sector_snapshots', 'market_snapshots', 'todos', 'capital_flow_cache']:
+    try:
+        tcols = [r[1] for r in t.execute(f"PRAGMA table_info({table})")]
+        mcols = [r[1] for r in m.execute(f"PRAGMA table_info({table})")]
+        print(f'{table}: third_cols={tcols}')
+        print(f'         main_cols={mcols}')
+        tcnt = t.execute(f"SELECT COUNT(*) FROM {table}").fetchone()[0]
+        mcnt = m.execute(f"SELECT COUNT(*) FROM {table}").fetchone()[0]
+        print(f'         third={tcnt} rows, main={mcnt} rows')
+    except Exception as e:
+        print(f'{table}: ERR {e}')
+    print()
+
+# sector_snapshots schema in both
+print('sector_snapshots sample (third):')
+for r in t.execute("SELECT * FROM sector_snapshots ORDER BY rowid DESC LIMIT 2"):
+    print(' ', str(r)[:200])
+print('sector_snapshots sample (main):')
+for r in m.execute("SELECT * FROM sector_snapshots ORDER BY rowid DESC LIMIT 2"):
+    print(' ', str(r)[:200])
+
+t.close()
+m.close()
@@ -0,0 +1,94 @@
+import sqlite3, os, shutil
+from datetime import datetime
+
+THIRD = '/home/hmo/.hermes/profiles/position-analyst/scripts/data/mofin.db'
+MAIN = '/home/hmo/MoFin/data/mofin.db'
+BACKUP = f'/home/hmo/MoFin/archive/third-db-backup-{datetime.now().strftime("%Y%m%d-%H%M")}'
+os.makedirs(BACKUP, exist_ok=True)
+shutil.copy(THIRD, os.path.join(BACKUP, 'mofin.db'))
+print('backup:', BACKUP)
+
+t = sqlite3.connect(THIRD, timeout=10)
+m = sqlite3.connect(MAIN, timeout=30)
+m.execute('PRAGMA busy_timeout=30000')
+
+# 1. sector_snapshots: dedup by (snapshot_id, name)
+existing = set()
+for r in m.execute("SELECT snapshot_id, name FROM sector_snapshots"):
+    existing.add((r[0], r[1]))
+print(f'main sector_snapshots keys: {len(existing)}')
+
+cols = ['snapshot_id', 'name', 'change_pct', 'up_count', 'down_count', 'net_inflow',
+        'lead_stock', 'lead_stock_change', 'volume', 'turnover']
+ins = 0
+skip = 0
+for r in t.execute(f"SELECT {', '.join(cols)} FROM sector_snapshots"):
+    if (r[0], r[1]) in existing:
+        skip += 1
+        continue
+    m.execute(f"INSERT INTO sector_snapshots ({', '.join(cols)}) VALUES ({','.join('?'*len(cols))})", r)
+    existing.add((r[0], r[1]))
+    ins += 1
+m.commit()
+print(f'sector_snapshots: inserted {ins}, skipped {skip}')
+
+# 2. market_snapshots: dedup by (timestamp, source)
+existing2 = set()
+for r in m.execute("SELECT timestamp, source FROM market_snapshots"):
+    existing2.add((r[0], r[1]))
+cols2 = ['timestamp', 'source', 'up_ratio', 'mood', 'created_at']
+ins2 = 0
+skip2 = 0
+for r in t.execute(f"SELECT {', '.join(cols2)} FROM market_snapshots"):
+    if (r[0], r[1]) in existing2:
+        skip2 += 1
+        continue
+    m.execute(f"INSERT INTO market_snapshots ({', '.join(cols2)}) VALUES ({','.join('?'*len(cols2))})", r)
+    existing2.add((r[0], r[1]))
+    ins2 += 1
+m.commit()
+print(f'market_snapshots: inserted {ins2}, skipped {skip2}')
+
+# 3. todos: third has 4 rows, main has 94 — check by title, insert missing (map to main schema)
+todos_t = t.execute("SELECT title, description, status, priority, source, fix_action, retry_count, note, created_at, updated_at FROM todos").fetchall()
+ins3 = 0
+for r in todos_t:
+    ex = m.execute("SELECT id FROM todos WHERE title=?", (r[0],)).fetchone()
+    if ex:
+        continue
+    m.execute("INSERT INTO todos (title, description, status, priority, source, fix_action, retry_count, note, created_at, updated_at) VALUES (?,?,?,?,?,?,?,?,?,?)", r)
+    ins3 += 1
+m.commit()
+print(f'todos: inserted {ins3}, skipped {len(todos_t)-ins3}')
+
+# 4. capital_flow_cache: keep whichever is newer
+t_row = t.execute("SELECT cache_json, updated_at FROM capital_flow_cache ORDER BY id DESC LIMIT 1").fetchone()
+m_row = m.execute("SELECT cache_json, updated_at FROM capital_flow_cache ORDER BY id DESC LIMIT 1").fetchone()
+if t_row and (not m_row or (t_row[1] or '') > (m_row[1] or '')):
+    m.execute("DELETE FROM capital_flow_cache")
+    m.execute("INSERT INTO capital_flow_cache (cache_json, updated_at) VALUES (?,?)", t_row)
+    m.commit()
+    print('capital_flow_cache: replaced with third (newer)')
+else:
+    print('capital_flow_cache: main is newer/equal, kept')
+
+# verify
+total = m.execute("SELECT COUNT(*) FROM sector_snapshots").fetchone()[0]
+total2 = m.execute("SELECT COUNT(*) FROM market_snapshots").fetchone()[0]
+print(f'after merge: sector_snapshots={total}, market_snapshots={total2}')
+
+m.close()
+t.close()
+
+# 5. 删除第三库（连同 data 目录里的其他残留）
+third_dir = os.path.dirname(THIRD)
+trash = '/home/hmo/trashbox/third-db-retired-20260720'
+os.makedirs(trash, exist_ok=True)
+shutil.move(THIRD, os.path.join(trash, 'mofin.db'))
+print('third db moved to', trash)
+for f in os.listdir(third_dir):
+    src = os.path.join(third_dir, f)
+    if os.path.isfile(src):
+        shutil.move(src, os.path.join(trash, f))
+        print('  also moved:', f)
+print('DONE')
@@ -0,0 +1,32 @@
+import json, urllib.request
+
+msg = """[笑笑] price_events.json 已彻底退役 — 全系统只认 DB
+
+按你的指示，所有和 price_events.json 相关的代码已全部改完：
+
+【代码修改（6个文件）】
+1. price_monitor.py — record_event 只写 DB，EVENTS_PATH/load_events/save_events 全部删除
+2. strategy_feedback.py — 价格事件只从 DB 读，JSON fallback 删除
+3. system_health_check.py — 文件检查清单移除 price_events.json，事件统计只读 DB
+4. mo_config.py — 删除无调用的 price_events_path 死属性
+5. mofin_health.py — 新鲜度检查改回读 DB 表（现在它是权威了）
+6. mofin_db.py — 修了一个更大的坑 ⬇️
+
+【顺藤摸出的更大的坑：三个数据库】
+mofin_db.py 的 DB_PATH 原来是相对路径（__file__.parent/data），导致每个硬链接副本解析到不同的库：主库、web-dashboard 库、还有 profile 本地一个几乎全空的第三库。健康检查之前读到 0 条事件就是因为读到了那个空库。已改为绝对路径 /home/hmo/MoFin/data，4 份 mofin_db 副本已统一硬链接。
+
+【验证结果】
+✅ record_event 只写 DB，JSON 不再被重建
+✅ system_health_check：历史事件 6353 条 / 今日事件 2965 条（之前误报 0/0）
+✅ strategy_feedback / price_monitor 完整跑通
+✅ price_events.json 已归档 trashbox（数据早已全量回填 DB）
+
+JSON 时代结束。"""
+
+payload = json.dumps({"to": "hmo@yoin.fun", "body": msg, "type": "chat"}).encode()
+req = urllib.request.Request("http://127.0.0.1:5805/", data=payload,
+                             headers={"Content-Type": "application/json"})
+try:
+    print("XMPP:", urllib.request.urlopen(req, timeout=10).read().decode()[:80])
+except Exception as e:
+    print("XMPP fail:", e)
@@ -0,0 +1,38 @@
+import json, urllib.request
+
+msg = """[笑笑] 系统冗余全面审计完成 — 报告已生成
+
+按你的要求只审计不动手。报告位置：
+📄 本地 vault: daily/调研笔记/系统冗余审计-20260720.md
+📄 246 Obsidian: daily/系统冗余审计-20260720.md
+
+核心发现（9 类问题）：
+
+🔴 高危×3：
+1. 31 个同名 .py 文件内容分叉（deploy vs MoFin/scripts vs MoFin根 vs .hermes/scripts）——今天"修了还报错"的病根就是这个
+2. 数据库碎片化：除主库外还有一个"第三库"（含 17340 行 sector 数据需合并）+ 3 个废弃小库
+3. 僵尸进程：xmpp_xiaoguo_bot 以 root 跑了 8 天吃 2.5GB 内存；xiaoguo-tunnel 对着不可达的 Mac Mini 循环重连
+
+🟡 中危×4：
+4. 三套 dashboard（8899/5803/9090）+ 两套 health check + 两套 auto_heal
+5. 小果生态残留：2 进程 + 4 脚本 + 2 服务 + 2 cron（扫描死了但其余还活着）
+6. 26 个 >14 天的遗留 JSON 文件
+7. projects/ 与根目录两对重复目录（1.4G + 2.2M）
+
+🟢 低危×2：
+8. 10 个被禁用的 cron 任务
+9. 18 个旧日志 + 旧 .bak 文件
+
+报告末尾有 7 个需要你拍板的问题（Q1-Q7），最关键的两个：
+- Q1: xiaoguo gateway (8645) 的"压缩策略"还在用吗？
+- Q2: xiaoguo-quick-scan（今天还在跑）和 market_scanner 什么关系？
+
+你看完报告定方向，我再按决定执行。"""
+
+payload = json.dumps({"to": "hmo@yoin.fun", "body": msg, "type": "chat"}).encode()
+req = urllib.request.Request("http://127.0.0.1:5805/", data=payload,
+                             headers={"Content-Type": "application/json"})
+try:
+    print("XMPP:", urllib.request.urlopen(req, timeout=10).read().decode()[:80])
+except Exception as e:
+    print("XMPP fail:", e)
@@ -0,0 +1,52 @@
+import json, shutil, os
+from datetime import datetime
+
+# 1. 删小果相关 cron job（quick-scan / tunnel-watchdog / 情感分析 / 独立扫描）
+REMOVE_JOBS = {'xiaoguo-quick-scan', 'xiaoguo-tunnel-watchdog', '小果情感分析', '小果独立扫描'}
+for jf in ['/home/hmo/.hermes/profiles/position-analyst/cron/jobs.json',
+           '/home/hmo/.hermes/cron/jobs.json']:
+    shutil.copy(jf, jf + '.bak-20260720-xiaoguo')
+    d = json.load(open(jf))
+    is_list = isinstance(d, list)
+    jobs = d if is_list else d.get('jobs', [])
+    kept = [j for j in jobs if j.get('name') not in REMOVE_JOBS]
+    removed = [j.get('name') for j in jobs if j.get('name') in REMOVE_JOBS]
+    if removed:
+        print(f'{jf}: removed {removed}')
+        if is_list:
+            json.dump(kept, open(jf, 'w'), ensure_ascii=False, indent=2)
+        else:
+            d['jobs'] = kept
+            json.dump(d, open(jf, 'w'), ensure_ascii=False, indent=2)
+    else:
+        print(f'{jf}: nothing to remove')
+
+# 2. 归档小果脚本和数据文件
+ARCHIVE = '/home/hmo/MoFin/archive/xiaoguo-retired-20260720'
+os.makedirs(ARCHIVE, exist_ok=True)
+candidates = [
+    '/home/hmo/MoFin/deploy/profile-scripts/xiaoguo_scanner.py',
+    '/home/hmo/MoFin/deploy/profile-scripts/xiaoguo_news_processor.py',
+    '/home/hmo/MoFin/deploy/profile-scripts/xiaoguo_sentiment_bridge.py',
+    '/home/hmo/MoFin/deploy/profile-scripts/xiaoguo_signal_consumer.py',
+    '/home/hmo/MoFin/scripts/inject_xiaoguo_insight.py',
+    '/home/hmo/web-dashboard/data/xiaoguo_insights.json',
+    '/home/hmo/web-dashboard/data/xiaoguo_sentiment.json',
+    '/home/hmo/MoFin/data/xiaoguo_insights.json',
+    '/home/hmo/MoFin/data/xiaoguo_sentiment.json',
+]
+for f in candidates:
+    if os.path.exists(f):
+        shutil.move(f, os.path.join(ARCHIVE, os.path.basename(f)))
+        print('archived:', f)
+
+# 3. 处理 profile scripts 里残留的 xiaoguo 硬链接
+import glob
+for f in glob.glob('/home/hmo/.hermes/profiles/position-analyst/scripts/xiaoguo_*.py'):
+    os.unlink(f)
+    print('unlinked:', f)
+
+# 4. 检查 default profile scripts 的 xiaoguo 文件
+for f in glob.glob('/home/hmo/.hermes/scripts/xiaoguo_*.py') + glob.glob('/home/hmo/.hermes/scripts/xiaoguo_*.sh'):
+    print('default profile xiaoguo file:', f)
+print('DONE')
@@ -0,0 +1,18 @@
+import shutil, os, glob
+ARCHIVE = '/home/hmo/MoFin/archive/xiaoguo-retired-20260720'
+os.makedirs(ARCHIVE, exist_ok=True)
+for f in ['/home/hmo/.hermes/scripts/xiaoguo_quick_scan.py',
+          '/home/hmo/.hermes/scripts/xiaoguo_scanner.py',
+          '/home/hmo/.hermes/scripts/xiaoguo_news_processor.py',
+          '/home/hmo/.hermes/scripts/xiaoguo_tunnel_watchdog.sh',
+          '/home/hmo/xmpp_xiaoguo_bot.py']:
+    if os.path.exists(f):
+        shutil.move(f, os.path.join(ARCHIVE, os.path.basename(f)))
+        print('archived:', f)
+# xiaoguo profile dir
+p = '/home/hmo/.hermes/profiles/xiaoguo'
+if os.path.isdir(p):
+    print('xiaoguo profile dir exists:', os.listdir(p)[:8])
+else:
+    print('no xiaoguo profile dir')
+print('DONE')
@@ -0,0 +1,106 @@
+import os, shutil, glob
+from datetime import datetime
+
+ARCHIVE = f'/home/hmo/MoFin/archive/dedup-20260720-{datetime.now().strftime("%H%M")}'
+os.makedirs(ARCHIVE, exist_ok=True)
+
+def hardlink_to(src, targets):
+    """把 targets 全部硬链接到 src（内容统一为 src）"""
+    done = []
+    for t in targets:
+        if not os.path.exists(t) and not os.path.islink(t):
+            continue
+        if os.path.isfile(t) and os.path.samefile(src, t):
+            continue
+        try:
+            if os.path.islink(t) or os.path.isfile(t):
+                os.unlink(t)
+            os.link(src, t)
+            done.append(t)
+        except Exception as e:
+            print(f'  LINK FAIL {t}: {e}')
+    return done
+
+def archive(path, note=''):
+    if os.path.exists(path) or os.path.islink(path):
+        dst = os.path.join(ARCHIVE, os.path.basename(path))
+        if os.path.exists(dst):
+            dst = dst + '.' + datetime.now().strftime('%H%M%S')
+        shutil.move(path, dst)
+        print(f'  archived: {path} {note}')
+
+# ── A. LIBRARY 类：root 是活的（cron 经 sys.path 导入），统一所有副本到 root ──
+LIBRARY = ['strategy_lifecycle.py', 'strategy_tree.py', 'technical_analysis.py',
+           'multi_timeframe.py', 'mofin_news.py']
+print('=== A. LIBRARY 统一到 MoFin root ===')
+for f in LIBRARY:
+    root = f'/home/hmo/MoFin/{f}'
+    if not os.path.exists(root):
+        print(f'  {f}: root 不存在，跳过!')
+        continue
+    targets = [
+        f'/home/hmo/MoFin/deploy/profile-scripts/{f}',
+        f'/home/hmo/.hermes/profiles/position-analyst/scripts/{f}',
+        f'/home/hmo/MoFin/scripts/{f}',
+        f'/home/hmo/web-dashboard/{f}',
+    ]
+    done = hardlink_to(root, targets)
+    print(f'  {f}: unified {len(done)} copies -> root')
+
+# ── B. CRON 类：deploy 是唯一源，归档其他副本 ──
+CRON_FILES = ['advice_reconciliation.py', 'branch_scanner.py', 'collect_evaluation_data.py',
+              'cron_to_xmpp.py', 'market_insight.py', 'memory_guardian.py', 'mofin_health.py',
+              'premarket_full_review.py', 'promote_candidates.py', 'prune_branches.py',
+              'strategy_evaluator.py', 'system_health_check.py', 'system_audit.py']
+print('\n=== B. CRON 类归档非 deploy 副本 ===')
+for f in CRON_FILES:
+    for loc in ['/home/hmo/MoFin', '/home/hmo/MoFin/scripts',
+                '/home/hmo/web-dashboard', '/home/hmo/.hermes/scripts']:
+        p = os.path.join(loc, f)
+        # 不碰 deploy 和 pa/scripts（硬链接到 deploy）
+        if os.path.exists(p) or os.path.islink(p):
+            # 跳过与 deploy 同 inode 的（那些是合法硬链接）
+            deploy_p = f'/home/hmo/MoFin/deploy/profile-scripts/{f}'
+            if os.path.exists(deploy_p) and os.path.isfile(p) and not os.path.islink(p):
+                try:
+                    if os.path.samefile(p, deploy_p):
+                        continue
+                except Exception:
+                    pass
+            archive(p, f'(cron file {f})')
+
+# ── C. market_watch / market_screener 特殊：crontab 正在跑 root 旧版 ──
+print('\n=== C. market_watch/market_screener 统一 ===')
+# 暂时只统一内容到 deploy（deploy 是 hermes cron 维护的版本），crontab 稍后改指向
+for f in ['market_watch.py', 'market_screener.py']:
+    deploy_p = f'/home/hmo/MoFin/deploy/profile-scripts/{f}'
+    if os.path.exists(deploy_p):
+        done = hardlink_to(deploy_p, [f'/home/hmo/MoFin/{f}', f'/home/hmo/MoFin/scripts/{f}'])
+        print(f'  {f}: unified {len(done)} -> deploy (crontab 路径仍有效但内容已统一)')
+
+# ── D. 待查类 ──
+print('\n=== D. 待查类 ===')
+# server.py: web-dashboard + root 是活的 dashboard，归档 deploy 旧拷贝
+archive('/home/hmo/MoFin/deploy/profile-scripts/server.py', '(live=web-dashboard hardlink)')
+# xmpp_agent_core.py: deploy/bot 是活的
+for p in ['/home/hmo/MoFin/xmpp_agent_core.py', '/home/hmo/MoFin/scripts/xmpp_agent_core.py',
+          '/home/hmo/.hermes/scripts/xmpp_agent_core.py', '/home/hmo/web-dashboard/xmpp_agent_core.py']:
+    archive(p, '(live=deploy/bot)')
+# 手动工具类：deploy 为准，归档 root 旧副本
+for f in ['mo_provider.py', 'mofin_collect.py', 'stock_profile.py', 'stock_sector_enrich.py',
+          'trend_detector.py', 'bulk_strategy_regenerate.py', 'strategy_feedback.py']:
+    deploy_p = f'/home/hmo/MoFin/deploy/profile-scripts/{f}'
+    if os.path.exists(deploy_p):
+        done = hardlink_to(deploy_p, [f'/home/hmo/MoFin/{f}', f'/home/hmo/MoFin/scripts/{f}'])
+        print(f'  {f}: unified {len(done)} -> deploy')
+    else:
+        for loc in ['/home/hmo/MoFin', '/home/hmo/MoFin/scripts']:
+            archive(os.path.join(loc, f), '(no deploy version)')
+
+print('\n=== 验证：关键 import 测试 ===')
+import subprocess
+r = subprocess.run(
+    "cd /home/hmo/.hermes/profiles/position-analyst/scripts && python3 -c \"import sys; sys.path.insert(0,'/home/hmo/MoFin'); import strategy_lifecycle, technical_analysis, multi_timeframe, mofin_news, strategy_tree; print('imports OK', strategy_lifecycle.__file__)\"",
+    shell=True, capture_output=True, text=True, timeout=30)
+print(r.stdout.strip() or r.stderr.strip()[:200])
+print('\nARCHIVE dir:', ARCHIVE)
@@ -0,0 +1,49 @@
+{
+  "module": "hygiene",
+  "version": "1.0",
+  "purpose": "系统卫生审计——防止冗余/废弃/僵尸问题复发的每周自动检查（红线6-10 的 enforcement）",
+  "human_help": {
+    "title": "系统卫生审计说明",
+    "description": [
+      "每周一 07:30 自动运行 system_hygiene_audit.py，检查六类系统卫生问题并推送 XMPP 报告。",
+      "这是 2026-07-20 冗余事件（三个数据库、31个分叉文件、小果僵尸进程）后建立的防复发机制。"
+    ],
+    "usage": [
+      "手动运行: cd /home/hmo/.hermes/profiles/position-analyst/scripts && python3 system_hygiene_audit.py",
+      "查看报告: /home/hmo/MoFin/gateway/logs/hygiene_report.json"
+    ],
+    "troubleshooting": [
+      "分叉副本(diverged_copy): 同名 .py 在不同位置内容不同 → 归档旧副本或硬链接到权威版",
+      "断裂硬链接(broken_hardlink): deploy 和 profile scripts 内容不一致 → 跑 sync_profile_scripts.sh",
+      "孤儿文件(orphan_data_file): 生产数据目录 >14 天的文件 → 归档到 archive/（红线8）"
+    ]
+  },
+  "ai_spec": {
+    "apis": [],
+    "dependencies": [
+      "deploy/profile-scripts/system_hygiene_audit.py",
+      "deploy/profile-scripts/sync_profile_scripts.sh",
+      "docs/dev-spec.md 红线 6-10"
+    ],
+    "constraints": [
+      "检查项: 分叉副本/断裂硬链接/僵尸进程/孤儿文件/死cron/DB新鲜度",
+      "docker 容器内进程豁免僵尸检测",
+      "报告写入 gateway/logs/hygiene_report.json，有问题推 XMPP(:5805)",
+      "活文件注册表 LIVE_DATA_FILES 需随新数据文件引入而更新"
+    ],
+    "must_not": [
+      "禁止只报告不修复——发现问题必须在当轮处理或创建 TODO",
+      "禁止将生产数据目录下 >14 天文件默认视为正常"
+    ],
+    "tests": [
+      {"id": "H1", "name": "人为制造分叉副本，下一轮审计能发现"},
+      {"id": "H2", "name": "sync 断裂后审计报 broken_hardlink"},
+      {"id": "H3", "name": "全部检查通过时 status=ok 且不推 XMPP"}
+    ],
+    "related_files": [
+      "deploy/profile-scripts/system_hygiene_audit.py",
+      "deploy/profile-scripts/sync_profile_scripts.sh",
+      "gateway/logs/hygiene_report.json"
+    ]
+  }
+}