JF Garbage Overflow Inspection for Openclaw

1. Install the Python dependency

pip install requests

2. Configure JF API credentials

Set the required environment variables before running an inspection:

export JF_UUID="your-user-uuid"
export JF_APP_KEY="your-app-key"
export JF_APP_SECRET="your-app-secret"
export JF_MOVE_CARD="2"
export JF_ENDPOINT="api-cn.jftechws.com"

Use the regional endpoint appropriate for the deployment: api-cn.jftechws.com, api-as.jftechws.com, api-eu.jftechws.com, or api-na.jftechws.com. JF_ENDPOINT is optional and defaults to the China endpoint.

3. Prepare device prerequisites

Ensure each JF monitoring device has completed network configuration and binding, is online, and is available through the dependent jf-open-pro-device-list and jf-open-pro-device-status skills.

4. Configure inspection cameras

Create config/cameras.json with the devices and analysis settings:

{
  "cameras": [
    {
      "sn": "device-serial-number",
      "name": "Building A Floor 1 Garbage Bin",
      "password": "device-password",
      "location": "Building A Lobby"
    }
  ],
  "settings": {
    "overflow_threshold": "moderate",
    "default_channel": 0
  }
}

5. Run a batch inspection

PYTHONIOENCODING=utf-8 python {skill_dir}/scripts/capture_and_download.py \
  --action inspect-batch \
  --config {skill_dir}/config/cameras.json \
  --output-dir {workspace}/captures/{YYYYMMDD}/ \
  --json

6. Run a single-device inspection

PYTHONIOENCODING=utf-8 python {skill_dir}/scripts/capture_and_download.py \
  --action inspect-single \
  --device-sn "device-serial-number" \
  --device-name "Building A Garbage Bin" \
  --device-location "Building A Lobby" \
  --output-dir {workspace}/captures/{YYYYMMDD}/ \
  --json

The agent then reads successful image files, analyzes them, generates the inspection report, and presents the captured images. Keep device passwords and API secrets protected because they are used for authentication and token retrieval.

Configuration file

Generated files and result data

• scripts/capture_and_download.py handles capture, download, authentication, and JSON output.
• scripts/crypto.py provides signing, encryption, and timestamp utilities imported by the capture script.
• {workspace}/captures/{YYYYMMDD}/ stores downloaded inspection images grouped by inspection date.
• The capture script returns JSON entries containing device metadata, success or failure status, image file paths, and error information.
• Successful image paths are read for visual analysis and supplied to present_files for user review.

Report taxonomy

Reports include inspection time, total devices, successful and failed counts, device name, serial number, location, status, analysis details, failure reasons, and a prioritized summary.