google-agents-cli-workflow Skill for Claude: Build ADK Agents on Google Cloud

The entry skill of Google's agents-cli: it walks Claude through designing, building, testing, deploying and monitoring an ADK agent.

Installs: 532.3K; on skills.sh
Stars: 6.3K; on GitHub
Files: 6; besides SKILL.md
Updated: Oct 2026; last repo commit

google-agents-cli-workflow is the front door of Google's agents-cli skills. It runs Claude through eight phases: understand, study recipes, scaffold, build, evaluate, deploy, publish and observe. Each phase tells Claude which sibling skill to load next.

It starts with a design chat, asking one question at a time and writing a spec to .agents-cli-spec.md that you must approve. It tells Claude to resist common shortcuts, to treat evaluation as the most important phase and never to deploy without your approval.

Good for

Developers building agents on Google Cloud with ADK
Teams who want one repeatable path from idea to deployed agent
Coding-agent users who want Claude to follow rules, not improvise

Install google-agents-cli-workflow

Claude Code

Copy this prompt.

Install the google-agents-cli-workflow skill in Claude Code. Run this command:

uvx google-agents-cli setup

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires Python 3.11+, uv, Node.js.

Pick one route. setup installs the agents-cli program and the skills into your coding agent. The README says that with the skills-only route your coding agent handles the rest. The seven skills in the suite point to each other, so take the whole set.

From the project README. Source.

What's in the skill

When it loads

This skill should be used when the user wants to "develop an agent", "build an agent using ADK", "run the agent locally", "debug agent code", "test an agent", "deploy an agent", "publish an agent", "monitor an agent", or needs the ADK (Agent Development Kit) development lifecycle and coding guidelines. Entrypoint for building ADK agents. Always active — provides the full workflow (scaffold, build, evaluate, deploy, publish, observe), code preservation rules, model selection guidance, and troubleshooting steps for ADK or any agent development.

Sections in SKILL.md

Session Continuity & Skill Cross-References
Setup
uv command not found
Product name mapping
Phase 0: Understand
Phase 1: Study Reference Recipes
Phase 2: Scaffold (if needed)
Phase 3: Build and Implement
Provision recipe infrastructure (if you adapted one)
Phase 4: Evaluate
Phase 5: Deploy
Phase 6: Publish (optional)
Phase 7: Observe
Common Shortcuts to Resist
Principle 1: Code Preservation & Isolation
Principle 2: Execution Best Practices

Key facts

Author: Google
License: Apache-2.0 (stated in SKILL.md)
GitHub stars: 6,276 for the whole google/agents-cli repo (Nov 2, 2026)
Last commit: Oct 27, 2026
What installs: SKILL.md plus 6 bundled files (guides and examples); no scripts in the file list
Needs: agents-cli (uv tool install google-agents-cli); the other six skills in the suite
Part of: The google-agents-cli plugin in google/agents-cli
skills.sh audits: Gen Agent Trust Hub pass, Socket pass, Snyk pass
Works in: Antigravity CLI, Claude Code, Codex and other coding agents (README)

What it includes

Design chat, one question at a time, then an approved spec
Writes .agents-cli-spec.md from a template after you approve it.
Phase 0: Understand
Loads adk-code and checks for a matching worked recipe
Read · Phase 1: Study recipes
agents-cli scaffold create or enhance
Runs code · Phase 2: Scaffold
Writes agent code and smoke-tests with agents-cli run
Also suggests agents-cli playground for manual testing.
Runs code · Phase 3: Build
agents-cli eval run, expecting 5-10+ iterations
Says not to write unit tests that assert on LLM output.
Runs code · Phase 4: Evaluate
agents-cli deploy, only with explicit human approval
Runs code · Phase 5: Deploy
Registers with Gemini and sets up tracing
Runs code · Phases 6 and 7: Publish, Observe
Code preservation, model selection and shortcuts to resist
Operational guidelines

From the official docs

Permissions

Medium risk

Claude installs and runs the agents-cli program, writes project files and can deploy to Google Cloud. The skill says never to deploy without your approval, so keep that gate.

What it can do

Read your data
Create and change things
Delete things
Read-only mode

Watch out for

It can lead to a real Google Cloud deploy
Phase 5 runs agents-cli deploy. The skill says never to deploy without explicit human approval, so say no unless you mean it.
It installs a CLI with uv
The skill runs uv tool install google-agents-cli if the program is missing.
It creates files in your project
Scaffolding and the spec file add folders and files to the directory you are in.

Technical details

Scopes	Installs and runs the agents-cli program Creates project files and writes agent code Can deploy to Google Cloud when you approve

Questions

What does google-agents-cli-workflow do?

It is the entry skill of agents-cli. It gives Claude a phased plan for building an ADK agent: design, scaffold, build, evaluate, deploy, publish and observe.

How do I install it?

Use the install box at the top of this page. The setup command installs the agents-cli program and all seven skills in the suite, which is what you want. The README says that with the skills-only route your coding agent handles the rest. You need Python 3.11+, uv and Node.js first.

Do I need the other six skills?

Yes, it hands off to them at each phase. The seven skills in the suite cross-reference each other, so install the whole set.

Does it need an API key?

The SKILL.md names no API key. It needs the agents-cli program and a Google Cloud sign-in, which agents-cli login handles; the README says login also supports AI Studio.

Is it safe?

It runs CLI commands and can deploy to Google Cloud, but it tells Claude never to deploy without your approval. skills.sh lists Gen Agent Trust Hub, Socket and Snyk as pass.

Why does it make me answer questions first?

Phase 0 is a design dialogue that confirms what you are building before any scaffolding. It writes a spec you approve, and says not to proceed without it.

Does it work in the Claude app, Codex or Cursor?

The README names Antigravity CLI, Claude Code and Codex, and says any other coding agent works too. It does not name the Claude desktop or web app or Cursor.

Information

updated Nov 2, 2026

Repository: google/agents-cli
Folder: skills/google-agents-cli-workflow

google/agents-cli skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 3

google/agents-cli README · github.com
google-agents-cli-workflow SKILL.md · github.com
google-agents-cli-workflow on skills.sh · skills.sh