desktop-operator

name: desktop_operator description: Use repo scripts to inspect the desktop, move or click the mouse, and drive the managed OpenClaw browser.

Desktop Operator

Use this skill when the task requires local desktop inspection or controlled browser actions from this repo.

Repo root:

/Users/infatoshi/openclaw-course

Preferred path

Prefer the managed OpenClaw browser for web tasks because it is more deterministic than raw desktop clicks.

Examples:

• python3 scripts/browser_ctl.py start
• python3 scripts/browser_ctl.py open https://example.com --wait-load load
• python3 scripts/browser_ctl.py snapshot --limit 40
• python3 scripts/browser_ctl.py screenshot --copy artifacts/browser-latest.jpg

Desktop inspection and control

Take a screenshot before clicking if the target is not already obvious.

Examples:

• python3 scripts/desktop_ctl.py mouse-pos
• python3 scripts/desktop_ctl.py shot
• python3 scripts/desktop_ctl.py shot --region 1500 0 1200 1000
• python3 scripts/desktop_ctl.py click 1600 300
• python3 scripts/desktop_ctl.py type "hello world"
• python3 scripts/desktop_ctl.py key enter

Safety rules

• Read the latest screenshot before clicking on an uncertain target.
• Do not type or reveal secrets without explicit confirmation.
• Avoid destructive desktop actions unless the user clearly asked for them.
• Prefer browser refs and browser snapshots over coordinate clicks whenever possible.