Gemini CLI Adds Subagents with Separate Context Windows

Melvin Vivas · X post · 2026-04-16 · Open on X

Topics: AI Agents, Tool Use & MCP, AI Dev Tools & Productivity, Prompt & Context Engineering · Level: intermediate

Summary

The creator shares the news that Gemini CLI now supports subagents. Each subagent gets its own context window, custom system instructions and a chosen set of tools. This lets you build specialized agents, keep the main agent's context clean and run work in parallel.

Key points

• Gemini CLI now supports subagents.
• Each subagent has its own separate context window.
• Each subagent can have custom system instructions.
• Each subagent gets a curated set of tools.
• Use cases: specialized expert agents, keeping the main agent focused with a clean context, and handing work to parallel agents.

Resources mentioned

• Gemini CLI · tool · github.com · free
  Google's open-source terminal AI agent, which the creator says has been discontinued.
  Also in: Google Retires Gemini CLI, Launches Antigravity CLI (Melvin Vivas on X · notes), Google Shuts Down Gemini CLI and Firebase Studio (Melvin Vivas on X · notes), Gemini CLI sunset for individual tiers: migrate to Antigravity CLI (Melvin Vivas on X · notes), RepoPrompt 2.0 Agent: Context Builder + Oracle for Code Review and Planning (Melvin Vivas on X · notes)

Try this

• If you use Gemini CLI, try creating specialized subagents to keep the main agent's context clean.
• Set up a few specialized Gemini CLI subagents (for example a reviewer and a tester) and hand them tasks in parallel.