gemini-cli

Orchestrate Gemini runtime agents, tools, and extensions from the command line.

1|Updated Jan 22, 2026

Search Tags:#tooling #nodejs #extensions #sandbox #command-line #gemini-cli #execution-runtime

Ready to run

npx skills add https://github.com/guardian-intelligence/apm2 --skill gemini-cli-guardian-intelligence

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: gemini-cli
Source: https://github.com/guardian-intelligence/apm2/tree/main/documents/skills/gemini-cli
Command: npx skills add https://github.com/guardian-intelligence/apm2 --skill gemini-cli-guardian-intelligence

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Gemini CLI provides a structured, scriptable interface to interact with the Gemini runtime, enabling developers and operators to orchestrate AI agents, tools, and extensions from the command line.

Core Features & Use Cases

Interactive mode for exploratory work and debugging in a TTY.
Headless mode for automation and CI pipelines with deterministic outputs.
Extension and tool mediation to compose complex AI workflows safely.
Sandbox-aware execution and workspace management for reproducible runs.
Use Case: Integrate Gemini in a data science project to run a sequence of agent tasks and capture tool results in a replayable log.

Quick Start

Install Node.js (≥20) and install dependencies, then run the production entry bin.gemini to start the CLI in your project. For a quick look, run bin.gemini --help to see available commands.

Frequently Asked Questions about gemini-cli

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate Gemini runtime orchestration from the command line?▼

Automate Gemini runtime orchestration from the command line by running the CLI in headless mode, which provides deterministic outputs for CI pipelines. You configure execution via a CLI config with policy and extension support to manage agent tasks.

Does the Gemini CLI require a specific Node.js version and workspace layout?▼

The Gemini CLI requires Node.js ≥20 and a workspace monorepo layout. It uses esbuild bundling for production, ensuring sandbox-aware execution and reproducible runs within your configured project structure.

What is the difference between interactive and headless mode for command-line AI workflows?▼

Interactive mode supports exploratory work and debugging in a TTY, while headless mode enables automation and CI pipelines with deterministic outputs. Both modes handle workspace management, extension handling, and tool mediation during AI executions.

How do I compose complex AI workflows safely using command-line tool mediation?▼

Compose complex AI workflows safely by using the CLI's extension and tool mediation features. Configuration via a CLI config with policy and extension support allows you to sequence agent tasks and capture tool results in a replayable log.

Can I capture and replay tool results from agent tasks in a data science project?▼

You can capture tool results in a replayable log by running a sequence of agent tasks through the CLI. Sandbox-aware execution and workspace management ensure that these data science project runs remain reproducible.

When should I use sandbox-aware execution and workspace management for command-line AI agents?▼

Use sandbox-aware execution and workspace management when you need reproducible runs from your command-line AI agents. This approach isolates execution environments, ensuring deterministic outputs for automation and safe tool mediation.