Capture windows application
Skill NoLegitHere/capture-windows-application-mcp/skills/capture-windows-application
Use the capture-windows MCP to launch Windows executables, discover visible desktop windows, and capture application screenshots. Use when an agent needs to inspect or verify a native Windows app window, capture a running process by PID or title, or take multiple frames from a launched executable during local Windows UI validation.From its SKILL.md
npx -y skills add NoLegitHere/capture-windows-application-mcp --skill capture-windows-applicationAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
3.0 KB, 576 tokens by cl100k_base, as published. Nobody here has run it
Capture Windows Application
Use the capture-windows MCP for native Windows desktop screenshots. Prefer it for local .exe windows that browser tools cannot inspect.
Load The MCP
Use the configured MCP server named capture-windows. Harnesses may prefix MCP tool names differently, but this server exposes:
list_windows: discover visible desktop windows and their process IDs.capture_window: capture an already-running visible window byprocess_idor bywindow_title.run_and_capture: launch an executable, wait for its window, capture one or more screenshots, and optionally close it.
If these tools are unavailable, ask the harness to configure or enable this repo's local MCP server before claiming the window was inspected.
Choose A Tool
- Use
list_windowsfirst when the target app is already open or the title is uncertain. - Use
capture_windowwithprocess_idfromlist_windowswhenever possible. Usewindow_titleas a fallback or as a filter inside the selected process. - Use
run_and_capturewhen the task should launch a specific.exe. Pass a fullexecutable_path; addarguments,window_title, and a largerwait_mswhen startup or title matching needs it. - Use
capture_countandcapture_interval_msfor a short visual sequence. The MCP returns the last successful capture inline and reports saved paths for the capture set. - Set
close_afteronly when the launched or reused process should be terminated after capture.
Capture Workflow
For an open app:
- List windows.
- Select the visible window by process name, title, and PID.
- Capture by PID.
- Inspect the inline image or the saved PNG path before claiming a visual result.
For an app to launch:
- Resolve the executable path from the repo, build output, or known Windows path.
- Call
run_and_capture. - Increase
wait_msif the process starts before its window becomes visible. - Add
window_titlewhen one executable can own multiple visible windows. - Inspect the returned image before reporting visual verification.
Guardrails
- This MCP captures visible Windows desktop windows. It is not a DOM inspector and cannot validate hidden or headless UI.
- Prefer a user-provided or workspace-local
save_pathwhen the screenshot artifact should stay with the task. Without it, screenshots are saved under the MCP server's.screenshotsfolder. - Treat
run_and_captureas potentially reusing an already-running process with the same executable. Reviewclose_afterbefore enabling it. - If capture fails because no window appears, use
list_windows, inspect the actual title and PID, and retry with a better title filter or longer wait.
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.