[Bug] MCP tool read timeout is hardcoded to 60 seconds and not configurable

Environment

• OS: Linux/Windows
• Eclipse Platform Version: 4.38
• Plugin Version: GitHub Copilot 0.19.0

Describe the bug
The Copilot for Eclipse plugin enforces a 60-second read timeout on MCP (Model Context Protocol) tool calls. This timeout is hardcoded and cannot be configured by users or MCP server implementations.
In VS Code, this issue does not occur - likely due to a higher default timeout value. We request the Eclipse plugin to match VS Code’s behavior.

Minimal Reproducible Example
MCP Server (Python with FastMCP): python from fastmcp import FastMCP import asyncio mcp = FastMCP(“Example MCP Server”) @mcp.tool async def [long_running_build()](url) -> dict: “”“Simulates a build/compile operation that takes longer than 60 seconds.”“” await asyncio.sleep(90) # Simulate 90-second operation return {“success”: True, “message”: “Operation completed successfully”}

To Reproduce
Steps to reproduce the behavior:

1. Start the MCP server above
2. Configure Copilot for Eclipse to use this MCP server
3. Invoke the long_running_task tool via Copilot chat
4. Observe timeout error after 60 seconds, even though the task would complete at 90 seconds

Expected behavior
Match VS Code behavior (where this timeout issue does not occur) or Make timeout configurable via settings

No repository file or test is named. Start from the MCP tool call initiated through Copilot chat and reproduce the timeout with the provided 90-second Python example; done means the Eclipse plugin no longer fails at 60 seconds and either matches VS Code behavior or exposes a configurable timeout.