---

name: meshtastic-detection description: Receive DETECTION_SENSOR_APP alerts from Meshtastic LoRa devices via USB. When the remote sensor GPIO triggers (preset target detected), store the event and alert the user immediately. metadata: {"openclaw": {"os": ["darwin", "linux"], "requires": {"anyBins": ["python3.12", "python3.11", "python3.10", "python3"]}, "emoji": "📡"}}

Meshtastic Detection Skill

Receive detection sensor alerts from a remote Meshtastic device over LoRa. When the remote device's GPIO pin triggers (preset target detected), the event is stored locally and requires immediate user notification via feishu.

Prerequisites

• Meshtastic-compatible hardware connected via USB (RAK4631, T-Beam, Heltec, etc.)
• Python 3.10+ with meshtastic and pypubsub packages (venv at {baseDir}/venv)
• usb_receiver.py daemon running
• Quick setup: cd {baseDir} && ./setup.sh
• Detailed guide: {baseDir}/references/SETUP.md

Architecture

┌──────────────────────────────────────────────────────────────┐
│                     USB Receiver Daemon                       │
├──────────────────────────────────────────────────────────────┤
│  LISTEN:  DETECTION_SENSOR_APP only (GPIO trigger events)    │
│  STORE:   data/sensor_data.jsonl (append per detection)      │
│  LATEST:  data/latest.json (most recent detection)           │
└──────────────────────────────────────────────────────────────┘

┌─────────────┐     USB      ┌──────────────┐
│  LoRa Node  │◄────────────►│ usb_receiver │
│  (Radio)    │              │   daemon     │
└─────────────┘              └──────┬───────┘
                                    │
                    ┌───────────────┼───────────────┐
                    │               │               │
                    ▼               ▼               ▼
          sensor_cli.py     event_monitor.py   OpenClaw cron
          (query data)      (check alerts)     (feishu alert)

Quick Reference

Run the Receiver

cd {baseDir}
source venv/bin/activate
python scripts/usb_receiver.py --port /dev/cu.usbmodem1CDBD4A896441

Check for New Alerts

cd {baseDir}
./venv/bin/python scripts/event_monitor.py

Every DETECTION_SENSOR_APP record = high-priority alert. Output:

{
  "alerts": [{"priority": "high", "sender": "!1dd29c50", "text": "alert detected", "received_at": "...", "channel": "ch0", "portnum": "DETECTION_SENSOR_APP"}],
  "summary": "🚨 3 new detection alert(s) from 3 record(s)",
  "alert_count": 3,
  "new_records": 3
}

Query Historical Data

python scripts/sensor_cli.py latest
python scripts/sensor_cli.py stats --since 24h
python scripts/sensor_cli.py query --since 1h

Data Format

Each record in data/sensor_data.jsonl:

{"received_at": "2026-03-04T11:07:06+00:00", "sender": "!1dd29c50", "channel": "ch0", "portnum": "DETECTION_SENSOR_APP", "data": {"type": "detection", "text": "alert detected"}}

Only DETECTION_SENSOR_APP messages are captured. This portnum means the remote sensor's GPIO pin was triggered — a preset target has been detected. Every detection event requires immediate user alert.

All other message types (TEXT_MESSAGE_APP, telemetry, position, etc.) are ignored.

Log Rotation

sensor_data.jsonl is automatically rotated at 5 MB (keeps 2 archive files, total max ~15 MB). Rotation is transparent — event_monitor auto-resets offset, sensor_cli reads across archives.

Monitoring & Alerts

Cron Job (Active)

The cron job runs event_monitor.py every 60 seconds and delivers alerts to feishu:

# Check status
openclaw cron list

# View run history
openclaw cron runs --id <job-id>

# Manual test
openclaw cron run <job-id>

# Edit config
openclaw cron edit <id> --timeout-seconds 60 --to <feishu-open-id>

Cron message template (for reference):

Run this command and report the output:
cd {baseDir} && ./venv/bin/python scripts/event_monitor.py
— If alert_count > 0, tell me how many alerts, the latest sender and time.
  If alert_count is 0, reply: 暂无新告警。

Key settings:

• timeoutSeconds: 60 (agent needs ~20-40s)
• channel: feishu
• delivery.to: ou_16c6dc8bda8ac97abfd0194568edee59

Alert Behavior

All DETECTION_SENSOR_APP events are treated as high priority. No rule configuration needed — every detection triggers an immediate alert. The alert message includes:

• Sender device ID
• Detection text (from remote sensor config)
• Timestamp

Configuration

Edit CONFIG.md to customize:

• Serial port — USB device path
• Notification channel — feishu (configured in OpenClaw)

Common Conversation Patterns

User asks about recent detections:

"What was detected in the last hour?"

Run: cd {baseDir} && ./venv/bin/python scripts/sensor_cli.py query --since 1h

User asks for statistics:

"Give me a summary of detections today"

Run: cd {baseDir} && ./venv/bin/python scripts/sensor_cli.py stats --since 24h

User asks about system status:

"Is the sensor still working?"

Run: cd {baseDir} && ./venv/bin/python scripts/sensor_cli.py status

Files

{baseDir}/
├── SKILL.md               # This file (agent instructions + metadata)
├── CONFIG.md              # User configuration
├── setup.sh               # One-click setup
├── scripts/
│   ├── usb_receiver.py    # USB serial daemon (DETECTION_SENSOR_APP only)
│   ├── event_monitor.py   # Incremental alert monitor
│   └── sensor_cli.py      # Query CLI
├── data/
│   ├── sensor_data.jsonl  # Detection records (auto-rotated at 5 MB)
│   ├── latest.json        # Most recent detection
│   └── monitor_state.json # Monitor byte offset + seen hashes
└── references/
    └── SETUP.md           # Detailed installation guide

Troubleshooting

"No records found"

• Check that usb_receiver.py is running
• Verify USB device: ls /dev/cu.usb*

"Resource temporarily unavailable"

• Only one process can use the serial port. Check: lsof /dev/cu.usbmodem*

Receiver connects but no data appears

• The receiver only captures DETECTION_SENSOR_APP messages (other types are ignored)
• Run with --debug to see all packets: python scripts/usb_receiver.py --port ... --debug
• Verify the remote device is on the same channel and frequency
• Confirm the remote device has Detection Sensor Settings configured (GPIO pin monitoring)

Cron job times out or fails delivery

• Check: openclaw cron runs --id <job-id>
• Fix timeout: openclaw cron edit <id> --timeout-seconds 60
• Fix delivery: openclaw cron edit <id> --to <feishu-open-id>