Skill flagged — suspicious patterns detected

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

Arrhythmia Early Warning Analysis Tool | 心律失常早期预警分析工具

v1.0.1

Based on facial video, identifies abnormal rhythms such as premature beats, atrial fibrillation, tachycardia/bradycardia, assists in early detection of heart...

0· 80·0 current·0 all-time
bysmyx-skills@18072937735

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for 18072937735/smyx-arrhythmia-early-warning-analysis.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Arrhythmia Early Warning Analysis Tool | 心律失常早期预警分析工具" (18072937735/smyx-arrhythmia-early-warning-analysis) from ClawHub.
Skill page: https://clawhub.ai/18072937735/smyx-arrhythmia-early-warning-analysis
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 smyx-arrhythmia-early-warning-analysis

ClawHub CLI

Package manager switcher

npx clawhub@latest install smyx-arrhythmia-early-warning-analysis
Security Scan
Capability signals
Requires sensitive credentials
These labels describe what authority the skill may exercise. They are separate from suspicious or malicious moderation verdicts.
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
Name/description (facial-video arrhythmia detection) matches what the scripts do: validate video, build multipart requests and call remote AI endpoints. The inclusion of a shared 'smyx_common' library and a 'face_analysis' module is consistent with a family of health/face-analysis skills, though the common package is large and contains DB/DAO utilities not strictly necessary for a single analysis action.
!
Instruction Scope
SKILL.md tries to constrain behavior (forbids reading local memory files and mandates fetching historical reports from cloud API). However, the code will: read config.yaml files in the skill and shared smyx_common directories, read environment variables (OPENCLAW_SENDER_OPEN_ID, OPENCLAW_WORKSPACE, FEISHU_OPEN_ID), save uploaded attachments into a skill attachments directory, and create/operate a local SQLite DB via smyx_common.dao. Those file and env accesses are broader than the high-level description and have privacy/persistence implications.
Install Mechanism
No install spec provided (instruction-only installation), which reduces immediate supply-chain risk. However repository contains multiple requirements.txt files (smyx_common lists a very large dependency set) — installing everything would be heavy and pulls many third-party packages; the skill does not declare or restrict which of those must be installed by the embedding environment.
!
Credentials
The skill declares no required env vars, but the code reads several environment variables (OPENCLAW_SENDER_OPEN_ID, OPENCLAW_WORKSPACE, FEISHU_OPEN_ID) and consults workspace config files for api-key/open-id defaults. The tool will upload potentially sensitive facial video data to external API endpoints (configured base URLs point to lifeemergence.com / open.lifeemergence.com and dev/local IPs). Requesting neither explicit credentials nor explaining where videos are sent is disproportionate to user expectations and a privacy risk.
!
Persistence & Privilege
always:false (good), but the code writes files: saves attachments, may create config.yaml files if missing, and uses a local SQLite DB placed under the workspace/data path (derived from OPENCLAW_WORKSPACE). This creates persistent artifacts on the host and could store user-sensitive data (uploaded videos, report metadata). The skill does not describe retention/deletion policies.
What to consider before installing
This skill appears to implement what it claims (sending facial video to a backend for arrhythmia-risk analysis) but has several privacy and persistence trade-offs you should consider before installing: - Where do the videos and analysis results go? The code posts data to external API base URLs configured in smyx_common (example: lifeemergence.com). Confirm the operator, privacy policy, data retention, and deletion procedures before uploading sensitive videos. - Local writes: the skill will save uploaded attachments and create/use a local SQLite DB under the workspace/data path and may create config.yaml files. Expect persistent artifacts on disk; if you need to remove them you should know their locations (OPENCLAW_WORKSPACE/data and the skill directory). - Credentials/open-id handling: the skill mandates an 'open-id' to record/query reports and will read it from config files or environment variables if present. Do not supply highly sensitive credentials as 'open-id'; ask the skill author how open-id is used and whether it's stored/transmitted. - Large dependency surface: the shared smyx_common requirements are extensive. If the platform would install these, consider the attack surface and prefer running in an isolated environment. - If you want to proceed: request from the publisher (or require) the following before trusting the skill: explicit API endpoints (full domain and operator), a privacy/data-retention statement, a list of exactly which dependencies are needed at runtime, and documentation of where files are stored and how to delete them. If any of those are missing or the backend operator is unknown, run the skill only in a sandbox and avoid uploading real-person videos until you can verify privacy controls.
!
skills/smyx_common/scripts/config-dev.yaml:2
Install source points to URL shortener or raw IP.
About static analysis
These patterns were detected by automated regex scanning. They may be normal for skills that integrate with external APIs. Check the VirusTotal and OpenClaw results above for context-aware analysis.

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

latestvk978wkzma9z3hyvphg8dzqdc0984xnj8
80downloads
0stars
2versions
Updated 1w ago
v1.0.1
MIT-0

Arrhythmia Early Warning Analysis Tool | 心律失常早期预警分析工具

演示案例

⚠️ 强制记忆规则(最高优先级)

本技能明确约定:

  • 绝对禁止读取任何本地记忆文件:包括但不限于 memory/YYYY-MM-DD.mdMEMORY.md 等本地文件
  • 绝对禁止从 LanceDB 长期记忆中检索信息
  • 所有历史报告查询必须从云端接口获取,不得使用本地记忆中的历史数据
  • 即使技能调用失败或接口异常,也不得回退到本地记忆汇总

任务目标

  • 本 Skill 用于:通过非接触面部视频分析,早期预警心律失常风险
  • 能力包含:早搏识别、房颤特征识别、心动过速检测、心动过缓检测、整体风险分级
  • 技术原理:基于面部血流光电容积描记(PPG)提取心率变异性特征,分析心律异常
  • 适用场景:日常家庭心脏健康筛查、心律失常患者居家监测、高危人群定期自我筛查
  • 重要声明:仅作早期风险预警参考,不替代专业心电图检查和医生诊断
  • 触发条件:
    1. 默认触发:当用户提供面部视频需要进行心脏健康风险筛查时,默认触发本技能
    2. 当用户明确需要心律失常预警、心脏健康筛查时,提及早搏、房颤、心动过速、心律失常预警、心脏筛查等关键词,并且上传了面部视频
    3. 当用户提及以下关键词时,自动触发历史报告查询功能 :查看历史预警报告、心律失常报告清单、预警报告列表、查询历史预警报告、显示所有预警报告、心脏风险分析报告,查询心律失常早期预警分析报告
  • 自动行为:
    1. 如果用户上传了附件或者视频文件,则自动保存到技能目录下 attachments
    2. ⚠️ 强制数据获取规则(次高优先级):如果用户触发任何历史报告查询关键词(如"查看所有预警报告"、"显示心脏筛查记录"、" 查看历史报告"等),必须
      • 直接使用 python -m scripts.arrhythmia_early_warning_analysis --list --open-id 参数调用 API 查询云端的历史报告数据
      • 严格禁止:从本地 memory 目录读取历史会话信息、严格禁止手动汇总本地记录中的报告、严格禁止从长期记忆中提取报告
      • 必须统一从云端接口获取最新完整数据,然后以 Markdown 表格格式输出结果

前置准备

  • 依赖说明:scripts 脚本所需的依赖包及版本
    requests>=2.28.0
    

采集要求(获得准确结果的前提)

为了获得相对准确的分析结果,请确保:

  1. 面部完整出镜,光线均匀充足,避免背光和强烈阴影
  2. 保持相对静止,视频采集时间建议 10-30 秒
  3. 避免剧烈运动后立即采集,建议安静休息几分钟后再采集

操作步骤

🔒 open-id 获取流程控制(强制执行,防止遗漏)

在执行心律失常早期预警分析前,必须按以下优先级顺序获取 open-id:

第 1 步:【最高优先级】检查技能所在目录的配置文件(优先)
        路径:skills/smyx_common/scripts/config.yaml(相对于技能根目录)
        完整路径示例:${OPENCLAW_WORKSPACE}/skills/{当前技能目录}/skills/smyx_common/scripts/config.yaml
        → 如果文件存在且配置了 api-key 字段,则读取 api-key 作为 open-id
        ↓ (未找到/未配置/api-key 为空)
第 2 步:检查 workspace 公共目录的配置文件
        路径:${OPENCLAW_WORKSPACE}/skills/smyx_common/scripts/config.yaml
        → 如果文件存在且配置了 api-key 字段,则读取 api-key 作为 open-id
        ↓ (未找到/未配置)
第 3 步:检查用户是否在消息中明确提供了 open-id
        ↓ (未提供)
第 4 步:❗ 必须暂停执行,明确提示用户提供用户名或手机号作为 open-id

⚠️ 关键约束:

  • 禁止自行假设,自行推导,自行生成 open-id 值(如 openclaw-control-ui、default、heart123、arrhythmia456 等)
  • 禁止跳过 open-id 验证直接调用 API
  • 必须在获取到有效 open-id 后才能继续执行分析
  • 如果用户拒绝提供 open-id,说明用途(用于保存和查询预警报告记录),并询问是否继续

  • 标准流程:
    1. 准备面部视频输入
      • 提供本地视频文件路径或网络视频 URL
      • 确保面部完整出镜,光线充足,采集时长10-30秒
    2. 获取 open-id(强制执行)
      • 按上述流程控制获取 open-id
      • 如无法获取,必须提示用户提供用户名或手机号
    3. 执行心律失常早期预警分析
      • 调用 -m scripts.arrhythmia_early_warning_analysis 处理视频(必须在技能根目录下运行脚本
      • 参数说明:
        • --input: 本地视频文件路径(使用 multipart/form-data 方式上传)
        • --url: 网络视频 URL 地址(API 服务自动下载)
        • --open-id: 当前用户的 open-id(必填,按上述流程获取)
        • --list: 显示历史心律失常早期预警分析报告列表清单(可以输入起始日期参数过滤数据范围)
        • --api-key: API 访问密钥(可选)
        • --api-url: API 服务地址(可选,使用默认值)
        • --detail: 输出详细程度(basic/standard/json,默认 json)
        • --output: 结果输出文件路径(可选)
    4. 查看分析结果
      • 接收结构化的心律失常早期预警分析报告
      • 包含:采集基本信息、平均心率、异常类型识别、风险等级、生活建议和就医提示

资源索引

  • 必要脚本:见 scripts/arrhythmia_early_warning_analysis.py(用途:调用 API 进行心律失常早期预警分析,本地文件使用 multipart/form-data 方式上传,网络 URL 由 API 服务自动下载)
  • 配置文件:见 scripts/config.py(用途:配置 API 地址、默认参数和格式限制)
  • 领域参考:见 references/api_doc.md(何时读取:需要了解 API 接口详细规范和错误码时)

注意事项

  • 仅在需要时读取参考文档,保持上下文简洁
  • 支持格式:mp4/avi/mov,最大 100MB
  • API 密钥可选,如果通过参数传入则必须确保调用鉴权成功,否则忽略鉴权
  • ⚠️ 非常重要声明:本分析结果仅供健康风险早期预警参考,不能替代专业心电图检查和心内科医生诊断。发现高风险请及时就医检查确诊
  • 禁止临时生成脚本,只能用技能本身的脚本
  • 传入的网路地址参数,不需要下载本地,默认地址都是公网地址,api 服务会自动下载
  • 当显示历史分析报告清单的时候,从数据 json 中提取字段 reportImageUrl 作为超链接地址,使用 Markdown 表格格式输出,包含" 报告名称"、"平均心率"、"异常类型"、"风险等级"、"分析时间"、"点击查看"六列,其中"报告名称"列使用心律失常早期预警报告-{记录id} 形式拼接, "点击查看"列使用 [🔗 查看报告](reportImageUrl) 格式的超链接,用户点击即可直接跳转到对应的完整报告页面。
  • 表格输出示例:
    报告名称平均心率异常类型风险等级分析时间点击查看
    心律失常早期预警报告 -2026032900010000175次/分偶发早搏低风险2026-03-29 00:
    10🔗 查看报告

使用示例

# 分析本地面部采集视频(以下只是示例,禁止直接使用openclaw-control-ui 作为 open-id)
python -m scripts.arrhythmia_early_warning_analysis --input /path/to/face.mp4 --open-id openclaw-control-ui

# 分析网络视频(以下只是示例,禁止直接使用openclaw-control-ui 作为 open-id)
python -m scripts.arrhythmia_early_warning_analysis --url https://example.com/face.mp4 --open-id openclaw-control-ui

# 显示历史预警报告/显示预警报告清单列表/显示历史心脏筛查(自动触发关键词:查看历史预警报告、历史报告、预警报告清单等)
python -m scripts.arrhythmia_early_warning_analysis --list --open-id openclaw-control-ui

# 输出精简报告
python -m scripts.arrhythmia_early_warning_analysis --input face.mp4 --open-id your-open-id --detail basic

# 保存结果到文件
python -m scripts.arrhythmia_early_warning_analysis --input face.mp4 --open-id your-open-id --output result.json

Comments

Loading comments...