End-to-end agent workflow support
Helps a coding assistant create, evaluate, deploy, publish, and observe AI agents on Google Cloud, rather than forcing the user to learn every CLI and service manually.
agents-cli
agents-cli is a CLI and skills package that helps coding assistants build, evaluate, deploy, and publish AI agents on Google Cloud. It supports both standalone terminal use and integrations with coding agents such as Antigravity CLI, Claude Code, and Codex.
Overview
agents-cli is a CLI and skill pack for building AI agents on Google Cloud, specifically on Gemini Enterprise Agent Platform. It is designed to make an existing coding assistant better at the full agent lifecycle: scaffolding projects, writing ADK-based code, running evaluations, deploying infrastructure, and publishing agents.
The repository positions the tool as a way to work with agents end to end without having to learn each Google Cloud service in isolation. It supports coding assistants such as Antigravity CLI, Claude Code, and Codex, and it can also be used directly from the terminal. The README notes prerequisites of Python 3.11+, uv, and Node.js for setup, and it distinguishes between local development workflows and cloud deployment features.
Key capabilities
Broad CLI command set
Provides setup, login, scaffold, run, install, lint, eval, deploy, publish, infra, data-ingestion, info, and update commands for working with agent projects from the terminal.
Skill-based coding assistant guidance
Adds skills for workflow guidance, ADK Python API usage, project scaffolding, evaluation methodology, deployment, publishing, and observability.
Evaluation and optimization tooling
Supports evaluation flows such as generating traces from eval datasets, grading results, synthesizing multi-turn scenarios, comparing result files, analyzing failure modes, and optimizing prompts.
Deployment and registration workflows
Includes deployment-related commands and scaffolding for Agent Runtime, Cloud Run, GKE, CI/CD, secrets, and Gemini Enterprise registration.
Flexible usage modes
Can be used with Antigravity CLI, Claude Code, Codex, or other coding agents, and can also run standalone from the terminal.
Typical use cases
Create a new agent project
Start a new agent project with scaffolding, then have your coding assistant apply the ADK code patterns and workflow guidance needed to build the first version.
Evaluate and tune an agent
Run eval datasets, generate traces, grade outputs, compare result files, and analyze failure modes when you want to check how an agent behaves before shipping it.
Deploy an agent to Google Cloud
Set up deployment to Google Cloud, including Agent Runtime, Cloud Run, GKE, CI/CD, and secrets, when the project needs a cloud-ready release path.
Extend an existing agent project
Use scaffold enhance to add deployment, CI/CD, or RAG to an existing project instead of rebuilding it from scratch.
Publish and monitor an agent
Register an agent with Gemini Enterprise and configure observability through logging, Cloud Trace, or other integrations once the agent is ready for broader use.
Pros and Cons
Pros
- Covers multiple stages of the agent lifecycle in one tool: scaffold, code, eval, deploy, publish, and observability.
- Works with more than one coding assistant and can also run standalone from the command line.
- Includes evaluation-focused commands for generating traces, grading results, comparing runs, and analyzing failure modes.
- Supports both new projects and existing projects through scaffold create, enhance, and upgrade commands.
- Documents infrastructure and data workflows such as CI/CD setup, datastore provisioning, and data ingestion.
Cons
- Cloud deployment and publishing features depend on Google Cloud; the README says local development can be done without it, but deployment and cloud features require it.
- The project is a tool for coding agents rather than a full agent itself, so users still need another assistant or terminal-based workflow to operate it.
FAQ
Is agents-cli itself a coding assistant?
No. The README says agents-cli is a tool for coding agents, not a coding agent itself. It provides CLI commands and skills that help another coding assistant build, evaluate, and deploy agents.
Do I need Google Cloud to use it?
For local development tasks such as creating, running, and evaluating agents, the README says you do not need Google Cloud and can use an AI Studio API key to run Gemini locally with ADK. Cloud deployment features do require Google Cloud.
Which coding assistants does it work with?
The README lists support for Antigravity CLI, Claude Code, Codex, and any other coding agent. It also says the CLI can be used standalone from the terminal.
What kinds of workflows does it support?
The source describes commands for setup, scaffolding, evaluation, deployment, publishing, infrastructure provisioning, data ingestion, and observability. The exact workflow depends on whether you are starting a new project or adding deployment and CI/CD to an existing one.
How is agents-cli priced?
The pricing page shows GitHub offers a free plan and paid tiers, but the repository itself is open source on GitHub. The source provided here does not show a separate price for agents-cli itself.
Quick Facts
- Category
- Developer Tool
- Product type
- CLI and skills package
- Primary platform
- Google Cloud / Gemini Enterprise Agent Platform
- Supported assistants
- Antigravity CLI, Claude Code, Codex, and other coding agents
- Source domain
- github.com
- Prerequisites
- Python 3.11+, uv, and Node.js
Альтернативы agents-cli
Devin
Devin is an AI coding agent and software engineer that helps developers and engineering teams plan and execute complex software tasks. It is available through desktop, cloud, JetBrains, and CLI surfaces, with plans for individuals, teams, and enterprises.
ByteAsk
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repositories and verifies changes with the real compiler, debugger, sanitizers, and tests before showing a diff. It offers a free tier plus paid plans, with editor connectors and zero-retention handling described in the source.
CreateOS Sandbox
CreateOS Sandbox is an isolated compute environment for running code and agent workloads inside Firecracker micro-VMs. It is designed for workflows that need machine-level isolation, private networking between sandboxes, and programmatic control through SDK, CLI, or MCP.
Ghost
Ghost — терминальный AI-ассистент для чата, генерации кода и запуска задач в командной строке. Бесплатные модели, Linux, macOS, Windows, open source.
AakarDev AI
AakarDev AI helps teams manage AI provider access, project-level setups, logs, and analytics from one dashboard. It supports BYOK workflows and lists providers including OpenAI, Google Gemini, Anthropic, Groq, Mistral AI, and Perplexity AI.
Arduino VENTUNO Q
Arduino VENTUNO Q is an edge AI computer for AI and robotics applications. It combines AI inference and deterministic control on a single board and is designed to work with Arduino App Lab.