Built-in vs External Tools

Gemini CLI ships with a set of built-in tools that require zero configuration: read_file, write_file, edit, glob, search_file_content, run_shell_command, google_web_search, web_fetch, list_directory, codebase_investigator, save_memory, and write_todos. These tools are always available and always reliable. They form the foundation of every Gemini CLI workflow. External tools from MCP servers and extensions layer on top of this foundation.

The practical question is: when should you use a built-in tool versus an MCP server? Built-in tools handle file operations, search, and shell execution. MCP servers handle external systems — databases, APIs, cloud services, version control platforms. The dividing line is the boundary of your filesystem. If the tool needs to reach beyond your local machine, it is an MCP server. If it operates on local files and commands, built-in tools handle it.

BUILT-IN TOOLS (always available, zero config)
├── read_file, write_file, edit     → File operations
├── glob, list_directory             → File discovery
├── search_file_content              → Content search (ripgrep-style)
├── run_shell_command                → Shell execution
├── google_web_search, web_fetch     → Web access
├── codebase_investigator            → Deep code analysis
└── save_memory, write_todos         → Session state

EXTERNAL TOOLS (MCP servers, require config)
├── Database servers                 → Query, schema, migrations
├── GitHub/GitLab servers            → PRs, issues, code review
├── Cloud service servers            → Deploy, monitor, configure
└── Custom internal servers          → Your team's specific tools