Google Managed Agents add background runs

Google says Gemini API Managed Agents now support background execution, remote MCP servers, custom function calling, and credential refresh while preserving sandbox state across interactions. Why it matters: Hosted agent runtimes are moving toward durable workers instead of single request-response chats. Background runs and remote tool connections make it easier to design long-running tasks with status polling, retries, and explicit sandbox boundaries.

Try this: Prototype one low-risk background agent task with a short timeout, pollable status, and a narrow remote MCP/tool surface before moving any production workflow into a managed sandbox.