ワンクリックでManusで任意のスキルを実行

iot-monitor

スター0

フォーク0

更新日2026年3月10日 00:13

Dashboard and API layer for monitoring IoT sensors and edge devices from OpenClaw's full-stack environment. Connects to MQTT brokers, HTTP sensor APIs, or PicoClaw agent heartbeats. Renders a live web dashboard, stores readings in Supabase, and sends alerts via any configured channel (Telegram, Slack, WhatsApp). Adapted from PicoClaw's Go iot-monitor for OpenClaw's Node.js full-stack runtime.

インストール

Codex または Claude でインストールこの Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。

Manusで実行

ソース

mk-knight23

mk-knight23/AI-Agent-OpenClaw

GitHub リポジトリを開く Creator のリポジトリを見る

ダウンロード

Manusで実行

iot-monitor

IoT monitoring dashboard — connect PicoClaw edge devices to OpenClaw's full-stack visibility layer.

Usage

@OpenClaw iot-monitor --connect mqtt://192.168.1.100:1883
@OpenClaw iot-monitor --connect picoclaw://home-server:8080    # PicoClaw agent API
@OpenClaw iot-monitor --connect http://sensor-api.local/metrics
@OpenClaw iot-monitor --dashboard                              # Launch live web UI

Architecture

OpenClaw acts as the orchestration layer while PicoClaw runs on the hardware:

PicoClaw (Raspberry Pi)  →  MQTT / HTTP  →  OpenClaw Dashboard  →  Supabase + Alerts

Features

Live Dashboard

React-based real-time dashboard (WebSocket updates)
Time-series charts per sensor (last 24h, 7d, 30d)
Alert history and acknowledgment workflow
Device status: online/offline/degraded

Data Ingestion

MQTT subscription: all topics via sensors/# wildcard
PicoClaw heartbeat API integration
HTTP polling (configurable interval)
Schema inference: auto-detects numeric sensors vs. categorical

Alert Rules

{
  "rules": [
    {"sensor": "temperature", "condition": ">75", "channel": "telegram", "severity": "high"},
    {"sensor": "disk_usage", "condition": ">90%", "channel": "slack", "severity": "critical"}
  ]
}

Storage

All readings stored in Supabase iot_readings table
Automatic data retention: 90 days (configurable)
Export: CSV, JSON, InfluxDB line protocol

Files Created

dashboard/                  # Next.js dashboard app
  pages/iot.tsx             # Main monitoring page
  components/SensorChart.tsx
config/iot-monitor.json     # Connection and alert config
MONITOR_SETUP.md            # Integration guide for PicoClaw devices

Philosophy

PicoClaw owns the hardware layer. OpenClaw owns the visibility layer. This skill bridges them: raw sensor data becomes actionable insights in your full-stack dashboard.

このリポジトリの他の Skills

同じリポジトリ

batch-processor

mk-knight23/AI-Agent-OpenClaw

High-throughput parallel processing of large datasets using Node.js streams and Claude AI. Accepts CSV, JSON, JSONL, or plain text. Processes in configurable chunk sizes with backpressure-aware Node.js Readable/Transform/Writable streams. Tracks progress, retries failures with exponential backoff, estimates cost before run. Adapted from ZeroClaw's Rayon-based batch processor for Node.js streaming architecture.

2026-03-100

security-scanner

mk-knight23/AI-Agent-OpenClaw

Full-stack security audit for OpenClaw's Node.js ecosystem: npm audit for CVEs, Snyk for supply chain, ESLint security plugin for code patterns, OWASP dependency check, and AI review of the 60-repo portfolio for systemic vulnerabilities. Generates SECURITY_REPORT.md. Adapted from ZeroClaw's comprehensive security-scanner, tuned for Node.js/TypeScript and multi-repo portfolios.

2026-03-100

add-gmail

mk-knight23/AI-Agent-OpenClaw

Adds Gmail monitoring to OpenClaw using the googleapis library. Creates OAuth2 credentials, polls INBOX for unread messages matching a trigger pattern, drafts AI replies, and sends after approval. Use when you want OpenClaw to monitor and respond to emails. Requires a Google Cloud project with Gmail API enabled and OAuth2 credentials.

2026-03-060

add-obsidian

mk-knight23/AI-Agent-OpenClaw

Mounts an Obsidian vault as a read-only knowledge source for OpenClaw. Registers a vault_search tool that performs full-text search across all .md files, returning file names and relevant excerpts. Use when you want OpenClaw to answer questions using your personal notes or team knowledge base. Requires the vault to be accessible as a local directory.

2026-03-060

add-slack

mk-knight23/AI-Agent-OpenClaw

Adds Slack integration to OpenClaw using the @slack/bolt framework. Creates a SlackChannel adapter with Socket Mode (no public URL required), handles app_mention events, routes messages through OpenClaw's skill dispatcher, and responds in-thread. Use when you want OpenClaw accessible from Slack workspaces. Requires a Slack app with Socket Mode enabled.

2026-03-060

add-supabase

mk-knight23/AI-Agent-OpenClaw

Adds Supabase as OpenClaw's persistent backend. Creates tables for conversation history, skill invocation logs, and agent memory using the Supabase JS client. Wires memory-read and memory-write into the tool registry so skills can persist state across sessions. Requires a Supabase project URL and anon key.

2026-03-060

name	iot-monitor
description	Dashboard and API layer for monitoring IoT sensors and edge devices from OpenClaw's full-stack environment. Connects to MQTT brokers, HTTP sensor APIs, or PicoClaw agent heartbeats. Renders a live web dashboard, stores readings in Supabase, and sends alerts via any configured channel (Telegram, Slack, WhatsApp). Adapted from PicoClaw's Go iot-monitor for OpenClaw's Node.js full-stack runtime.

iot-monitor

IoT monitoring dashboard — connect PicoClaw edge devices to OpenClaw's full-stack visibility layer.

Usage

@OpenClaw iot-monitor --connect mqtt://192.168.1.100:1883
@OpenClaw iot-monitor --connect picoclaw://home-server:8080    # PicoClaw agent API
@OpenClaw iot-monitor --connect http://sensor-api.local/metrics
@OpenClaw iot-monitor --dashboard                              # Launch live web UI

Architecture

OpenClaw acts as the orchestration layer while PicoClaw runs on the hardware:

PicoClaw (Raspberry Pi)  →  MQTT / HTTP  →  OpenClaw Dashboard  →  Supabase + Alerts

Features

Live Dashboard

React-based real-time dashboard (WebSocket updates)
Time-series charts per sensor (last 24h, 7d, 30d)
Alert history and acknowledgment workflow
Device status: online/offline/degraded

Data Ingestion

MQTT subscription: all topics via sensors/# wildcard
PicoClaw heartbeat API integration
HTTP polling (configurable interval)
Schema inference: auto-detects numeric sensors vs. categorical

Alert Rules

{
  "rules": [
    {"sensor": "temperature", "condition": ">75", "channel": "telegram", "severity": "high"},
    {"sensor": "disk_usage", "condition": ">90%", "channel": "slack", "severity": "critical"}
  ]
}

Storage

All readings stored in Supabase iot_readings table
Automatic data retention: 90 days (configurable)
Export: CSV, JSON, InfluxDB line protocol

Files Created

dashboard/                  # Next.js dashboard app
  pages/iot.tsx             # Main monitoring page
  components/SensorChart.tsx
config/iot-monitor.json     # Connection and alert config
MONITOR_SETUP.md            # Integration guide for PicoClaw devices

Philosophy

PicoClaw owns the hardware layer. OpenClaw owns the visibility layer. This skill bridges them: raw sensor data becomes actionable insights in your full-stack dashboard.