google-agents-cli

Scaffold, evaluate, and deploy ADK agents to Google Cloud's Gemini Enterprise Agent Platform.

4|Updated May 16, 2026
One-click install
npx skills add https://github.com/reason-machines/devtools-skills --skill google-agents-cli-reason-machines
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: google-agents-cli
Source: https://github.com/reason-machines/devtools-skills/tree/main/skills/google-agents-cli
Command: npx skills add https://github.com/reason-machines/devtools-skills --skill google-agents-cli-reason-machines

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Building production AI agents on Google Cloud involves many disconnected steps: scaffolding projects, writing ADK agent code, running evaluations, provisioning infrastructure, and publishing to Gemini Enterprise. This Skill consolidates the entire agent development lifecycle into a single CLI workflow. ## Core Features & Use Cases - Project Scaffolding & Local Development: Create ADK agent projects from templates (basic, RAG), run agents locally with streaming output, and lint code. - Evaluation & Deployment: Run evalsets with metrics and LLM-as-judge, deploy to Cloud Run, GKE, or Agent Runtime, and set up CI/CD pipelines for staging and production. - Use Case: You need to ship a RAG-powered support agent. Scaffold the project with the RAG template, add a retrieval tool backed by a Vertex AI datastore, run evaluations against your evalset, then deploy to Cloud Run and publish to Gemini Enterprise—all through agents-cli commands. ## Quick Start Ask the AI to scaffold a new ADK agent project called my-agent and run it locally with a test prompt.

Frequently Asked Questions about google-agents-cli

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

FAQPage Schema
How do I create a new ADK agent project with agents-cli?▼

Run agents-cli scaffold my-agent to create a new project, optionally passing --template basic or --template rag. Then run agents-cli install to set up dependencies and agents-cli run with a prompt to test locally.

How to deploy an ADK agent to Google Cloud Run?▼

Authenticate with agents-cli login, then run agents-cli deploy with an optional --config deploy.yaml file specifying target, region, and resources. You can also deploy to a specific environment using --env production.

Does agents-cli support RAG agents and data ingestion?▼

Yes, agents-cli provides a RAG scaffold template, an infra datastore command to provision a Vertex AI datastore, and a data-ingestion command that loads documents from Cloud Storage buckets or local directories.

Can I run agents-cli locally without a Google Cloud project?▼

Yes, set the GOOGLE_API_KEY environment variable with an AI Studio API key for local development. Google Cloud authentication via agents-cli login is only required for deployment, infrastructure provisioning, and publishing.

Why does agents-cli deployment fail with authentication errors?▼

Deployment failures usually come from expired or missing credentials. Run agents-cli login --status to check, re-authenticate with agents-cli login --force, or set GOOGLE_APPLICATION_CREDENTIALS to a service account key for CI/CD environments.

How do I run evaluations with custom metrics in ADK?▼

Define evalsets in YAML with test cases and metrics, then run agents-cli eval run with --metrics accuracy,latency,cost. For custom logic, subclass the ADK Metric class and pass instances to an Evaluator in Python.