pyautogui-skill

Desktop automation via PyAutoGUI. Use when: user needs to automate mouse/keyboard actions, GUI testing, click/type sequences, screen-based workflows, or repe...

This skill appears coherent for local desktop automation, but be aware of the inherent risks of any tool that controls your mouse and keyboard: only run these scripts from sources you trust, keep pyautogui.FAILSAFE enabled (moving the mouse to a corner aborts actions), and avoid running while sensitive windows (password managers, banking, terminals) are focused. Install Python packages in a virtualenv and verify package versions before installing. On macOS grant Accessibility permissions only to trusted Terminal/Python apps. If you plan to allow autonomous agent invocation of this skill, understand that it could simulate input without further prompts — consider limiting autonomous use if you have high-sensitivity workflows.

Type: OpenClaw Skill Name: pyautogui-skill Version: 1.0.0 The skill provides powerful desktop automation capabilities via PyAutoGUI, including keyboard/mouse control and screen capture, which are inherently high-risk behaviors for an AI agent. While the scripts (e.g., click_image.py, keyboard_shortcut.py, take_screenshot.py) are clearly aligned with the stated purpose and include safety fail-safes, they provide a direct mechanism for an agent to bypass UI-based security prompts or exfiltrate sensitive information via screenshots. No evidence of intentional malice, obfuscation, or data exfiltration logic was found in the code.