Google Launches agents-cli: Developer Tools for Building Enterprise Agents on Gemini Platform
Key Takeaways
- ▸agents-cli enables coding assistants to autonomously build and deploy enterprise agents on Google Cloud with minimal developer intervention
- ▸The tool provides modular skills covering the complete agent development lifecycle: creation, evaluation, deployment, and Gemini Enterprise registration
- ▸Platform-agnostic design allows integration with Gemini CLI, Claude Code, Codex, and other coding agents, broadening accessibility
Summary
Google has introduced agents-cli, a command-line interface and skills package designed to enable coding assistants like Gemini CLI, Claude Code, and Codex to build, deploy, and manage enterprise-grade AI agents. The tool provides developers with a comprehensive toolkit for the full agent lifecycle—from development and evaluation to deployment on Google Cloud infrastructure, without requiring manual management of multiple CLIs and services. agents-cli includes six specialized skills covering workflow management, Python SDK integration, project scaffolding, evaluation methodology, deployment, and observability, along with dedicated CLI commands for each stage of agent development. The platform works seamlessly with multiple coding assistants and supports local development via AI Studio API keys, with full cloud deployment capabilities available through Google Cloud integration.
- Supports local development without Google Cloud (via AI Studio API) while enabling full-featured cloud deployment when needed
Editorial Opinion
Google's agents-cli represents a thoughtful approach to democratizing enterprise agent development by empowering coding assistants themselves to handle infrastructure complexity. By abstracting away Google Cloud's operational details into AI-friendly skills, Google is reducing the friction for developers building sophisticated multi-agent systems—though the tool's tight coupling with Google Cloud services and reliance on evolving coding assistant capabilities may limit adoption among teams committed to cloud-agnostic architectures.
