Windows Control
Analysis
The skill is coherent and not visibly malicious, but it gives an agent broad ability to see and control the entire Windows desktop, so it deserves careful review before installation.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.
else:
# Search all visible windows
windows = desktop.windows()
...
# Click the first matching element
target = candidates[0]
target['control'].click()If no window is specified, this click tool searches every visible window and clicks the first matching UI element, which creates a real risk of acting in the wrong application or approving an unintended action.
Source: unknown Homepage: none
The artifact provides limited provenance for a tool that can control the desktop, even though the included code is visible and no remote installer is shown.
Checks whether tool use, credentials, dependencies, identity, account access, or inter-agent boundaries are broader than the stated purpose.
Full Windows desktop control. Mouse, keyboard, screenshots - interact with any Windows application like a human.
The skill is explicitly designed to act through the user's Windows desktop, so it can inherit whatever access the user already has in open local apps and browser sessions.
Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.
screenshot = pyautogui.screenshot() ... print(img_base64)
The screenshot tool captures the full screen and returns it to the agent context, which can include private messages, documents, passwords shown on screen, or other sensitive information.
