sentinel

What This Skill Does

The Sentinel skill transforms an ordinary Android smartphone into a sophisticated, privacy-focused Edge AI security station. By interfacing with the 'IP Webcam' application, it leverages the device's camera, sensor array, and flash to create a low-power, high-autonomy monitoring system. Unlike cloud-based security cameras, Sentinel processes all pixel-differential analysis locally on your OpenClaw workspace, ensuring that video frames are never transmitted to external servers unless an anomaly threshold is triggered. It acts as a bridge between raw hardware telemetry and intelligent logic, allowing users to automate security routines, trigger external notifications, or initiate remote video recording based on precise motion-detection logic.

Installation

To integrate Sentinel into your environment, follow these steps:

1. Ensure your smartphone is connected to the same network as your OpenClaw instance.
2. Install the 'IP Webcam' app by Pavel Khlebovich on your device and launch the local server.
3. Verify connectivity by accessing the provided IP address (e.g., 192.168.1.100:8080) in your browser.
4. In your terminal, run the following command: clawhub install openclaw/skills/skills/alihmaou/ipwebcam.
5. Configure your local sentinel_runner.sh to target your specific device IP, ensuring that threshold variables are set according to your environment's lighting conditions.

Use Cases

• Intelligent Perimeter Monitoring: Automate the monitoring of entrances by utilizing the autofocus snapshot function for high-quality evidence gathering when movement is detected.
• Battery-Aware Security: Utilize the /sensors.json endpoint to monitor battery levels, triggering automatic alerts or power-saving modes when the device drops below a 20% charge.
• Dynamic Lighting Environments: Adjust focus, white balance, and torch status via scriptable endpoints, enabling the system to transition from daylight monitoring to night-vision modes seamlessly.
• Automated Archiving: Trigger remote video recordings directly through the OpenClaw agent, storing clips locally on the Android device for later retrieval.

Example Prompts

1. "Sentinel, check the current battery status of the garage camera and start a 10-second recording if the level is above 30%."
2. "Update the security system to night mode: enable the torch, set white balance to cloudy, and adjust the motion threshold to 1500."
3. "List all recorded security videos from the front door camera and download the most recent file to my workspace."

Tips & Limitations

To maintain optimal performance, calibrate your motion thresholds frequently; use 2500 for high-light conditions and 1500 for low-light. Be aware that 'lens flare' from static lights at night can trigger false positives, so position your camera away from direct glare. The system is designed for local-first operations, making it highly frugal with token usage, as no data is processed by the AI unless a pixel-differential threshold is crossed. Always check Wi-Fi stability to prevent gaps in monitoring coverage, and verify that the 'IP Webcam' app is set to run in the background with appropriate battery optimizations disabled to ensure continuous uptime.