The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes powershell.exe to bridge the WSL Linux environment and the Windows host operating system to perform host-level tasks.
[PRIVILEGE_ESCALATION]: The capture.sh script executes the PowerShell component using the -ExecutionPolicy Bypass flag. This bypasses the host's configured script execution policies to run the unsigned screenshot.ps1 script.
[DATA_EXFILTRATION]: The skill captures the full content of the user's Windows desktop, including all connected monitors. This allows the AI agent to access and process any sensitive information visible on the screen, such as private documents, communications, or credentials.
[DYNAMIC_EXECUTION]: The screenshot.ps1 script uses the Add-Type command to compile and inject a C# snippet into the PowerShell process at runtime to manage DPI awareness.
[INDIRECT_PROMPT_INJECTION]: The skill enables the agent to process untrusted visual data from the desktop, which may contain malicious instructions designed to influence the AI's behavior.
Ingestion points: The read tool is used to process the PNG output of the capture script as instructed in SKILL.md.
Boundary markers: None; the instructions do not include delimiters or warnings to treat captured visual data as untrusted.
Capability inventory: The skill has host-level execution capabilities via powershell.exe.
Sanitization: None; visual data is passed directly to the agent's vision capabilities without filtering.

wsl-screenshot