ClawHub

WOLP LAN Power Control

Wake or shut down LAN devices by sending WOL-plus packets from the agent host. Use this when the user wants to power on a device with a raw Ethernet magic packet, or power off a device with a UDP magic packet, and they can provide the target MAC address plus the network interface or target IPv4 address.

Audits

Pass
ClawScanReview

Agentic behavior and permission review.

Static analysisPass

Pattern checks against bundled files.

VirusTotalPass

Multi-engine malware detections and file reputation.

Install

openclaw skills install wolp

wolp

Use this skill when the user wants the agent to control a device on the local network.

Supported operations:

  • wake: send a standard UDP Wake-on-LAN magic packet with Python
  • shutdown: send a UDP magic packet to a target IPv4 address
  • list: print the resolved device inventory

Use the bundled Python script:

  • scripts/wolp_power.py
  • assets/devices.json

Required inputs:

  • Wake:
    • target MAC address
    • optional broadcast IPv4 address, default 255.255.255.255
    • optional UDP port, default 9
  • Shutdown:
    • target MAC address
    • target IPv4 address
    • optional extra_data, default FF:FF:FF:FF:FF:FF
    • optional UDP port, default 9

Constraints:

  • shutdown uses a normal UDP socket and does not require root.
  • wake uses the Python package wakeonlan and does not require a compiled helper.
  • Install the dependency before sending wake packets:
python3 -m pip install wakeonlan
  • The wake subcommand will print a clear error if wakeonlan is missing.
  • shutdown requires IP connectivity to the target host and a compatible WOL-plus listener on the target machine.
  • Packet send confirms only local transmission, not that the remote machine actually changed power state.

Device inventory:

  • Store reusable devices in skill/wolp/assets/devices.json.
  • Successful non-dry-run wake and shutdown commands automatically write the resolved device info back to assets/devices.json.
  • If --device <name> is provided, that entry is updated in place; otherwise the script reuses an existing entry with the same MAC or creates a new device-<mac> entry.
  • Repeated operations on the same device keep refreshing that device's stored fields and the latest success metadata.
  • The file format is:
{
  "defaults": {
    "broadcast_ip": "255.255.255.255",
    "port": 9,
    "extra_data": "FF:FF:FF:FF:FF:FF"
  },
  "devices": {
    "nas": {
      "mac": "AA:BB:CC:DD:EE:FF",
      "host": "192.168.1.50",
      "broadcast_ip": "192.168.1.255"
    },
    "desktop": {
      "mac": "11:22:33:44:55:66",
      "host": "192.168.1.60",
      "extra_data": "12:34:56:78:9A:BC",
      "last_action": "shutdown",
      "last_success_at": "2026-03-21T00:00:00Z",
      "port": 9
    }
  }
}
  • Use list before sending if you need to inspect or verify stored devices.
  • CLI flags override inventory values.

Preferred commands:

python3 skill/wolp/scripts/wolp_power.py list
python3 skill/wolp/scripts/wolp_power.py wake --device nas
python3 skill/wolp/scripts/wolp_power.py shutdown --device nas
python3 skill/wolp/scripts/wolp_power.py wake --mac AA:BB:CC:DD:EE:FF
python3 skill/wolp/scripts/wolp_power.py wake --mac AA:BB:CC:DD:EE:FF --broadcast-ip 192.168.1.255 --port 9
python3 skill/wolp/scripts/wolp_power.py shutdown --host 192.168.1.50 --mac AA:BB:CC:DD:EE:FF --extra-data FF:FF:FF:FF:FF:FF --port 9

For safe previews or debugging, use --dry-run first:

python3 skill/wolp/scripts/wolp_power.py wake --device nas --dry-run
python3 skill/wolp/scripts/wolp_power.py shutdown --device nas --dry-run

Client install and config:

  • Project: https://github.com/leeyeel/WOL-plus
  • Releases: https://github.com/leeyeel/WOL-plus/releases
  • Client receives shutdown packets and serves the Web UI.
  • Default Web UI access:
    • URL: http://<client-ip>:2025
    • username: admin
    • password: admin123

Agent standard install procedure:

  1. Confirm the minimum missing inputs only:
    • target OS: Windows, Debian/Ubuntu, or RPM-based Linux
    • target architecture when relevant: amd64 or arm64/aarch64
    • whether the agent can install directly on the target machine or must only provide user instructions
    • target machine IP if the user wants Web UI verification
  2. Choose the install source:
    • prefer a matching package from Releases
    • prefer the Debian package when the agent can reach a Debian/Ubuntu host over SSH
    • only build from this repo when a needed Debian package is unavailable from Releases
  3. Install by platform:
    • Windows:
      • download installer_windows_amd64_v<version>.exe from Releases
      • if the agent cannot control the Windows desktop session, tell the user to run the installer manually
      • after installation, verify the service is running and open http://<windows-ip>:2025
    • Debian/Ubuntu: 
      sudo dpkg -i wolp-client_<version>_amd64.deb
sudo systemctl status wolp.service
    • RPM Linux: 
      sudo rpm -ivh wolp-client-<version>-1.x86_64.rpm
sudo systemctl status wolp.service
  4. Debian build fallback from this repo: 
    bash scripts/build-deb.sh amd64 0.0.0-dev
sudo dpkg -i release/client/wolp-client_0.0.0-dev_amd64.deb
sudo systemctl status wolp.service
  5. Verify the client after install:
    • confirm wolp.service is active
    • confirm the Web UI responds at http://<client-ip>:2025
    • tell the user to change the default password after first login
  6. Configure the client when the user wants shutdown support:
    • edit /usr/local/etc/wolp/wolp.json
    • set mac_address to the client machine MAC that should receive the shutdown packet
    • set interface to the active NIC name on the client machine
    • set extra_data to match the sender's --extra-data
    • set udp_port to match the sender's --port
    • set shutdown_delay, username, and password as requested
  7. Remember the fixed defaults and path layout:
    • binary: /usr/local/bin/wolp
    • config: /usr/local/etc/wolp/wolp.json
    • web UI: /usr/share/wolp/webui
    • service: wolp.service
    • default extra_data=FF:FF:FF:FF:FF:FF
    • default udp_port=9
    • default shutdown_delay=60
    • default HTTP UI port 2025
  8. Keep protocol roles clear:
    • sender-side inventory interface matters only for wake
    • receiver-side udp_port and extra_data matter only for shutdown

When reporting results or performing installs:

  • echo the resolved broadcast IP, host, UDP port, and normalized MAC values
  • for wake, report that the packet is sent through the wakeonlan Python package
  • state clearly whether the script performed a real send or a dry run
  • if the user did not provide enough data, ask only for the missing MAC, target IPv4 address, or wake broadcast IP when needed
  • if you install the client, report the package source, package path, config path, Web UI URL, and the exact extra_data and udp_port values you configured