Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Anti 996 Reminder

v1.2.0

温馨健康提醒技能,每日12:05(午间)和23:00(睡前)推送暖心内容,支持打卡攒积分,舒缓压力、劝导早睡。适用于微信/QQ通道。

0· 125·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for devioslang/anti-996-reminder.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Anti 996 Reminder" (devioslang/anti-996-reminder) from ClawHub.
Skill page: https://clawhub.ai/devioslang/anti-996-reminder
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install anti-996-reminder

ClawHub CLI

Package manager switcher

npx clawhub@latest install anti-996-reminder
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
!
Purpose & Capability
Name/description (scheduled noon/night reminders, check-in points) match the provided content files and a small send_noon.py helper, but the skill claims automatic processing of user replies (parsing replies that contain '睡', updating points.json, managing history/streak/total) while there is no handler code or channel webhook/dispatch implementation included to accept inbound messages and perform writes. That missing implementation is a material inconsistency.
Instruction Scope
SKILL.md instructs registering cron jobs that embed python3 -c commands which read JSON files under /root/.openclaw/workspace/skills/anti-996-reminder/contents/*.json and produce message text. Reading those local files is coherent with the purpose. However, the docs also describe read/write operations on points.json for check-ins but do not include code or instructions showing how inbound messages are captured and how writes are performed safely; the example cron messages also include concrete sample account/to IDs which should be replaced. The use of inline shell/python in cron messages is functional but grants the agent execution of small code snippets—this is expected for scheduling but worth noting.
Install Mechanism
No install spec; the package is instruction-only plus a tiny Python helper (send_noon.py) and static JSON content. Nothing is downloaded from external URLs and no archives are extracted. Low install risk.
Credentials
The skill declares no required environment variables or credentials. It references platform channel names and sample account/to IDs for delivery, but it does not request or include secrets. This is proportionate to the stated sending-only functionality. Be aware delivery relies on existing OpenClaw channel accounts which hold credentials elsewhere.
Persistence & Privilege
always:false and no special platform-wide privileges are requested. The skill registers scheduled jobs (expected for reminder functionality). It does not attempt to modify other skills or system-wide settings in the files provided.
What to consider before installing
This skill largely does what it says: scheduled reminders using local JSON content. However, before installing or enabling it, check the following: 1) The skill describes automated check-in processing (detect replies with '睡' and update points.json), but there is no code or webhook shown to receive inbound messages and perform those writes—confirm where reply handling will run and who/what will update points.json. 2) Replace the sample --account and --to values; do not leave the example IDs in production. 3) The cron examples embed inline python3 -c commands that will run on the agent host and read files under /root/.openclaw/... — ensure the agent runtime has appropriate file permissions and that these paths match your deployment. 4) The included send_noon.py only prints a random entry; it does not send messages or process replies—if you expect automatic check-in recording you must add a secure handler that validates incoming content and sanitizes inputs before writing to points.json. 5) Test in a sandbox channel first to verify messages, reply handling, and file writes behave as intended. If you want, provide the inbound-message handler code or explain how OpenClaw routes replies so I can reassess and raise confidence.

Like a lobster shell, security has layers — review code before you run it.

Runtime requirements

🌿 Clawdis
latestvk97f4jx4k253jfsjvphj7fwqth83nv7p
125downloads
0stars
4versions
Updated 1mo ago
v1.2.0
MIT-0

anti-996-reminder — 温馨健康提醒技能

功能概述

  • 每日 12:05 推送午间暖心提醒(舒缓压力、提醒休息)
  • 每日 23:00 推送睡前暖心提醒(劝导早睡、关怀放松)
  • 23点推送支持"打卡"回应,攒积分,记录连续打卡天数
  • 每周一 07:30 自动推送上周打卡周报(打卡天数 / 漏打卡天数)
  • 内容每次随机选取,避免重复

定时任务设置(CLI 手动注册)

本 skill 使用 OpenClaw 内置 cron 工具注册定时任务。以下是经过验证可正常推送的命令。

注意:isolated session 需要三个关键参数才能正常工作:

  • --light-context:跳过完整上下文加载,避免超时
  • --account <账号ID>:指定微信账号,否则投递失败
  • --best-effort-deliver:投递失败不阻塞任务

任务一:午间提醒(每天 12:05)

openclaw cron add \
  --name "anti-996-noon" \
  --cron "0 12 * * *" \
  --tz "Asia/Shanghai" \
  --session isolated \
  --light-context \
  --timeout-seconds 150 \
  --message "你是一个暖心的健康助手。现在是午间。用以下 bash 命令从午间内容池随机选一条内容:

python3 -c \"import json,random; c=json.load(open('/root/.openclaw/workspace/skills/anti-996-reminder/contents/noon.json')); item=random.choice(c); print(f'午安~现在是12:05,该休息一下啦。{item[\\\"text\\\"]}')\"

得到输出后,直接把这个完整句子作为你的回复输出即可(不要解释,不要加引号,不要加任何额外内容)。" \
  --announce \
  --account "8592acfc8006-im-bot" \
  --to "o9cq800M8K-wyrmql8S5MSqz9piM@im.wechat" \
  --channel openclaw-weixin \
  --best-effort-deliver

任务二:睡前提醒(每天 23:00)

openclaw cron add \
  --name "anti-996-night" \
  --cron "0 23 * * *" \
  --tz "Asia/Shanghai" \
  --session isolated \
  --light-context \
  --timeout-seconds 150 \
  --message "你是一个暖心的健康助手。现在是睡前时间。用以下 bash 命令从睡前内容池随机选一条内容:

python3 -c \"import json,random; c=json.load(open('/root/.openclaw/workspace/skills/anti-996-reminder/contents/night.json')); item=random.choice(c); print(f'夜深了~23点啦,放下手机,好好休息吧。{item[\\\"text\\\"]}\\n\\n回复【睡】打卡,今晚就赢1分~🌙')\"

得到输出后,直接把这个完整内容作为你的回复输出即可(不要解释,不要加引号,不要加任何额外内容)。" \
  --announce \
  --account "8592acfc8006-im-bot" \
  --to "o9cq800M8K-wyrmql8S5MSqz9piM@im.wechat" \
  --channel openclaw-weixin \
  --best-effort-deliver

⚠️ --account--to 需要替换为实际值。账号 ID 可通过 openclaw channels list 查看。

任务三:周报(每周一 07:30)

openclaw cron add \
  --name "anti-996-weekly" \
  --cron "0 7 * * 1" \
  --tz "Asia/Shanghai" \
  --session isolated \
  --light-context \
  --timeout-seconds 150 \
  --message "你是一个暖心的健康助手。现在是周一早上,来统计上周的打卡情况并发送周报。

用以下 python 命令计算:
python3 -c \"import json,datetime; data=json.load(open('/root/.openclaw/workspace/skills/anti-996-reminder/points.json')); history=data.get('history',[]); today=datetime.date.today(); lm=today-datetime.timedelta(days=today.weekday()+7); tm=today-datetime.timedelta(days=today.weekday()); wc=sum(1 for d in history if lm<=datetime.date.fromisoformat(d)<tm); print(f'{lm}~{(tm-datetime.timedelta(days=1))} 周报|打卡:{wc}/7|漏打卡:{7-wc}天|累计:{data.get(\\\"total\\\",0)}分|连续:{data.get(\\\"streak\\\",0)}天')\"

得到输出后,直接输出为一条温暖的消息即可,格式如下:

📋 上周睡眠周报(X月X日~X月X日)
✅ 打卡:X / 7 天
❌ 漏打卡:X 天
🌙 累计积分:X分
💪 连续打卡:X天

附加一句温暖的鼓励即可。" \
  --announce \
  --account "8592acfc8006-im-bot" \
  --to "o9cq800M8K-wyrmql8S5MSqz9piM@im.wechat" \
  --channel openclaw-weixin \
  --best-effort-deliver

打卡积分系统

积分规则

  • 每晚23点提醒后,用户回复""(含"睡"即可,不区分大小写)即为有效打卡
  • 每次打卡 +1 分,连续打卡天数同步记录
  • 断签则连续天数归零,但历史总分保留

积分查询

用户发送以下任意关键词时,触发积分查询:

  • "打卡"/"我的积分"/"积分"/"连续几天"/"打卡记录"

积分数据结构(存储于 points.json)

{
  "total": 0,
  "streak": 0,
  "lastCheckIn": "2026-03-24",
  "history": ["2026-03-20", "2026-03-21", "2026-03-22"]
}

打卡处理逻辑

当用户回复包含"睡"字时,执行以下步骤:

  1. 读取 points.json
  2. 检查 lastCheckIn 日期:
    • 如果是今天 → 回复"今晚已经打过卡啦~🌙 明天继续保持哦"
    • 如果是昨天 → streak + 1
    • 如果超过昨天 → streak 归 1(断签),附加一句"没关系,重新开始也是勇气~"
  3. total + 1,更新 lastCheckIn
  4. 组装回复:
    🌙 打卡成功!
    本月积分:X分
    连续打卡:X天
    
    [当 streak >= 3 时附加] 
    💪 已连续X天,真的很棒!
    
    [当 streak >= 7 时附加]
    🎉 一周啦!身体在悄悄感谢你~
    
    [当 streak >= 30 时附加]
    🏆 一个月!你是早睡达人了!
    
  5. 写回 points.json

内容池说明

  • noon.json:午间内容池(20条,职场/健康/心理/励志混合)
  • night.json:睡前内容池(21条,温暖治愈为主)

内容全程温暖正向,不依赖特定热点事件,不恐吓不说教。

文件结构

anti-996-reminder/
├── SKILL.md
├── contents/
│   ├── noon.json    # 午间内容池
│   └── night.json   # 睡前内容池
└── points.json      # 积分数据(自动创建)

📝 维护指南

内容池需要持续更新才能保持新鲜感。以下是维护节奏和规范。

基础规则

  • 随时可加:编辑 contents/noon.json / contents/night.json,无需重启,直接生效
  • 每周维护:建议每周打开内容池看一次,有感触就随手加几条
  • 不删旧条:已有内容除非有明显问题,否则保留,避免打破随机分布

内容格式规范

{
  "text": "内容文字...🌿",
  "tags": ["职场", "健康"]
}
  • text 必填,控制在 30-80 字
  • tags 可选,用于以后精细化推送(午间偏职场/健康,睡前偏心理/放松)
  • emoji 放在句尾,每条 1 个为宜
  • 内容全程温暖正向,不恐吓,不说教

标签分类参考

标签适用场景
职场加班、压力、工作与生活平衡
健康身体信号、饮食、运动
心理情绪、焦虑、放下
励志温暖鼓励、认可自己

更新节奏建议

每周:加 2-3 条新内容,观察是否出现"太硬""太说教"的内容并替换

节气/节日:可在内容池头部添加节日专属条,例如:

{ "text": "世界睡眠日,说晚安🌙 今天你睡够了吗?", "tags": ["健康"] }

维护检查清单

  • 每周打开 noon.jsonnight.json 各读一遍
  • 有新感悟/好句子,随手追加进去
  • 检查是否有内容重复(相似意思超过3条则合并)
  • 节日/热点前临时补充专属内容
  • 每月底检查积分数据 points.json 是否正常

禁止事项

  • ❌ 禁止插入任何政治、宗教敏感内容
  • ❌ 禁止恐吓型内容("不睡就会死"类)
  • ❌ 禁止硬广/推销内容
  • ❌ 禁止负面评价用户("你怎么又熬夜"类)

Comments

Loading comments...