ME News reports that on April 16 (UTC+8), according to monitoring by Beating, Google has officially launched the subagents feature for its open-source command-line programming tool, Gemini CLI. Each subagent has its own independent context window, system instructions, and set of tools. The main agent automatically delegates subtasks to the most suitable subagent when handling complex tasks, returning compressed summaries of the results without polluting the main session context. Gemini CLI comes with three built-in subagents: generalist (a general-purpose agent suited for high-tool-call tasks like bulk refactoring), cli_help (an expert guide for using Gemini CLI itself), and codebase_investigator (for exploring codebases, mapping architecture, and tracing bugs). Users can explicitly assign a subagent using the @agent syntax, such as `@codebase_investigator map the authentication flow`. Creating custom subagents is straightforward: simply place a Markdown file with YAML frontmatter metadata in the .gemini/agents/ directory, defining the role description, available tools, and system instructions. Subagents placed in ~/.gemini/agents/ are globally available, while those in a project directory are shared with the team alongside the code repository. Subagents can also be packaged into Gemini CLI extensions for distribution. Subagents support parallel execution, allowing multiple subagents to handle different subtasks simultaneously. However, Google cautions that parallel tasks involving code editing may result in write conflicts and should be used with care. The day before, Anthropic also released a redesigned desktop version of Claude Code, whose core feature is likewise multi-agent parallelism. AI programming tools are collectively shifting from a “single-agent conversation” model to an architecture of “main agent orchestration + expert agent execution,” driven primarily by context window management: as programming tasks grow more complex, cramming all intermediate steps into a single context leads to performance degradation and increased costs—splitting tasks into isolated agents has become the current engineering consensus. (Source: BlockBeats)
Gemini CLI Launches Subagent Feature for Custom AI Programming Teams
KuCoinFlashShare
Summary
On April 16 (UTC+8), Gemini CLI, Google’s open-source command-line tool, launched a subagent feature. Each subagent has its own context, instructions, and tools, with the main agent automatically assigning tasks. Built-in subagents include generalist, cli_help, and codebase_investigator. Users can create custom subagents using Markdown files. Subagents run in parallel, though code conflicts may occur. This update introduces new AI + crypto developments, demonstrating how on-chain news continues to evolve with advanced programming tools.
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information.
Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.