agents-cli v1.0.0: Official Google CLI to Scaffold, Evaluate & Deploy Production AI Agents (Claude Code / Cursor Compatible)

Google’s agents-cli v1.0.0 is a unified CLI tool that covers the entire AI agent development lifecycle. It bridges the gap between local prototyping and production deployment while providing a machine-readable interface that coding agents like Claude Code and Cursor can directly operate.

📑Table of Contents

1. What is agents-cli? Google’s Unified Agent Development Lifecycle Tool
2. Installation and Initial Setup
3. Workflow from Agent Creation to Local Evaluation
4. Production Deployment to Cloud Run / Agent Runtime
5. Real-World Use Cases and Considerations
6. Comparison of agents-cli with Existing Tools
7. Frequently Asked Questions (FAQ)
8. Summary

---

What is agents-cli? Google’s Unified Agent Development Lifecycle Tool

AI agent development often suffers from fragmentation between local creation and production deployment. agents-cli integrates with Google Cloud’s Agent Platform to solve this. According to the official Google Developers Blog, it is positioned as a dedicated CLI for coding agents. See the official documentation for more details.

---

Installation and Initial Setup

Installation completes with a single uvx command. Start with the following:

uvx google-agents-cli setup

Agent Mode is intended for AI agents such as Claude Code or Cursor to use directly, while Human Mode allows developers to operate from the terminal. Skills injection is also straightforward via this CLI. Check the official GitHub repository (https://github.com/google/agents-cli) for the latest information.

---

Workflow from Agent Creation to Local Evaluation

Creating an agent is simple with the create command.

agents-cli create finance-agent -y --deployment-target agent_runtime

After creation, use eval run or eval compare commands for local evaluation. The comparison feature allows parallel evaluation of multiple agents based on defined metrics. Specific evaluation harness details are covered in the official blog.

---

Production Deployment to Cloud Run / Agent Runtime

Transitioning to production is also a one-command operation.

agents-cli infra single-project
agents-cli deploy
agents-cli publish gemini-enterprise

It automates deployment to Cloud Run and Agent Runtime, with easy integration into CI/CD pipelines. High compatibility with existing Google Cloud projects is another advantage.

---

Real-World Use Cases and Considerations

An example travel expense reimbursement agent demonstrates practical workflows when combined with Human Intent. However, users should understand its optimization for specific use cases. Always refer to official sources for the latest support status.

---

Comparison of agents-cli with Existing Tools

Item	agents-cli	Traditional GCP CLI / SDK	Advantage
Agent-focused design	Dedicated	General-purpose	Direct operation by coding agents
Evaluation features	Built-in	Requires separate setup	Immediate comparison possible
Deployment	One command	Multiple steps	Time savings
Supported targets	Claude Code / Cursor	Manual	Optimized for AI developers

Source: Google Developers Blog (as of April 2026)

---

Frequently Asked Questions (FAQ)

Q: Is agents

cli free to use? A: Core features are open source and free. Google Cloud usage fees apply separately.

Q: Can it be used with tools other than Claude Code or Cursor?

It also supports Gemini CLI and other coding agents. Check the official documentation for details.

Q: How accurate is local evaluation?

The built-in evaluation harness enables reliable comparisons. Iterate improvements based on real operational results.

Q: Are deployment targets other than Cloud Run supported?

It supports publish to Agent Runtime and Gemini Enterprise as well.

Q: Is there compatibility with existing GCP projects?

The single-project mode allows integration with existing projects.

Q: What troubleshooting is available if errors occur?

Refer to the official GitHub Issues and documentation. Debugging in Human Mode is recommended.

---

Related articles:

• Google Agent Development Kit (ADK) Open Source Release — Production-Grade Multi-Agent Framework
• Google Gemini Managed Agents API GA — Execute Fully-Managed Agents via Single API Call in Secure Sandbox
• DiffusionGemma: 4x Faster Local Text Generation with Google Text Diffusion

Summary

agents-cli significantly improves the efficiency of AI agent development. Experience the consistent workflow from installation to deployment. Utilize the official repository and blog to stay updated with the latest information.