research

To automate long-form research while preserving AI tokens, you can spawn a background sub-agent using Gemini CLI to compile deep topic reports independently, saving detailed markdown results to a structured local directory.

The best way to run unattended background research is by launching a Gemini CLI sub-agent that operates independently over extended periods, allowing you to continue other work while it compiles a comprehensive markdown report.

The clarifying questions workflow prompts you with essential questions before starting ambiguous topic research, ensuring the spawned sub-agent focuses precisely on your requirements and generates a highly relevant structured report.

Yes, you can use this automated background research approach for product research and decision-making tasks, as it compiles thorough background knowledge and saves the output as a structured markdown report for easy reference.

To retrieve results after a deep research sub-agent finishes, the system notifies the main agent upon completion and provides a concise summary, while the full structured markdown report is saved automatically to a designated local path.

A limitation of using a spawned sub-agent for deep topic research is that it requires the Gemini CLI environment to function, and complex ambiguous topics may still require manual intervention to answer initial clarifying questions before research begins.