agents-cli

agents-cli is an open-source command-line tool published by Google under the Apache License 2.0, designed to give coding agents the skills and commands needed to build, evaluate, and deploy enterprise-grade AI agents on Google Cloud. It reached v1.0.0 in July 2026 and is part of the Gemini Enterprise Agent Platform. The project is available on PyPI as google-agents-cli and can be installed with a single uvx command.

What It Is

agents-cli sits between your coding assistant (such as Antigravity CLI, Claude Code, or OpenAI Codex) and the Google Cloud agent stack. Rather than replacing those coding agents, it equips them with structured "skills" — machine-readable instruction sets — that teach them how to scaffold, evaluate, deploy, and publish ADK-based agents end-to-end. The CLI also works standalone from any terminal without a coding agent present.

Agent Skills Architecture

The tool ships seven named skills that each cover a distinct phase of the agent-building lifecycle:

• google-agents-cli-workflow — development lifecycle, code preservation rules, model selection
• google-agents-cli-adk-code — ADK Python API: agents, tools, orchestration, callbacks, state
• google-agents-cli-scaffold — project scaffolding via create, enhance, and upgrade
• google-agents-cli-eval — evaluation methodology: metrics, datasets, LLM-as-judge, adaptive rubrics
• google-agents-cli-deploy — deployment to Agent Runtime, Cloud Run, GKE, CI/CD, and secrets management
• google-agents-cli-publish — Gemini Enterprise registration
• google-agents-cli-observability — Cloud Trace, logging, and third-party integrations

Skills can be installed independently via npx skills add google/agents-cli, letting a coding agent handle the rest of the setup.

CLI Command Surface

The full command set covers every stage from project creation to production:

• agents-cli setup — installs CLI and skills to coding agents
• agents-cli scaffold <name> — creates a new agent project from templates
• agents-cli scaffold enhance — adds deployment, CI/CD, or RAG to an existing project
• agents-cli run "prompt" — runs the agent with a single prompt for local testing
• agents-cli eval generate / grade / compare / analyze / optimize — full evaluation pipeline including LLM-as-judge grading and prompt auto-tuning
• agents-cli deploy — deploys to Google Cloud (Agent Runtime, Cloud Run, or GKE)
• agents-cli publish gemini-enterprise — registers the agent with Gemini Enterprise
• agents-cli infra cicd — provisions staging/production infrastructure and CI/CD pipelines

Deployment Model and Google Cloud Integration

For local development, agents-cli works with an AI Studio API key and runs Gemini via ADK without requiring a Google Cloud project. Cloud features — deployment, CI/CD, observability via Cloud Trace, and BigQuery analytics — require a Google Cloud project. The tool provisions infrastructure directly and manages secrets, making it a full-lifecycle operator rather than just a scaffolding helper. Deployed resources run in the user's own Google Cloud project under standard Google Cloud Service Terms.

Update: v1.0.0

The repository reached its first stable release, v1.0.0, on July 1, 2026, roughly three months after the project was created in April 2026. The GitHub repository shows 4,881 stars and 525 forks as of the data snapshot. The project is marked as a Pre-GA offering under Google Cloud's "Pre-GA Offerings Terms," meaning it is available as-is with potentially limited support while the platform matures toward general availability.