terminal_capture
Capture terminal screen output as PNG images to document or analyze command results and application states.
Instructions
Capture terminal screen as PNG screenshot
Args: session_id: ID of the terminal session
Returns: Dictionary with base64-encoded PNG image and metadata
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| session_id | Yes |
Output Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Implementation Reference
- The handler function decorated with @mcp.tool(), implementing the terminal_capture tool. It retrieves the XTermSession for the given session_id, calls capture_screenshot() on it, and returns the base64 image data and metadata.
@mcp.tool() async def terminal_capture(session_id: str) -> Dict[str, Any]: """Capture terminal screen as PNG screenshot Args: session_id: ID of the terminal session Returns: Dictionary with base64-encoded PNG image and metadata """ if session_id not in sessions: return {"status": "error", "error": f"Session {session_id} not found"} session = sessions[session_id] try: screenshot_data = await session.capture_screenshot() logger.info(f"Captured screenshot for session {session_id}") return { "session_id": session_id, "status": "captured", "image_data": screenshot_data["image_data"], "metadata": screenshot_data["metadata"], } except Exception as e: logger.error( f"Failed to capture screenshot for session {session_id}: {e}" ) return {"status": "error", "error": str(e)} - src/terminal_control_mcp/server.py:97-97 (registration)The @mcp.tool() decorator registers the terminal_capture function as an MCP tool.
@mcp.tool() - Docstring defining the input schema (session_id: str) and output schema (dict with session_id, status, image_data, metadata).
"""Capture terminal screen as PNG screenshot Args: session_id: ID of the terminal session Returns: Dictionary with base64-encoded PNG image and metadata