Google Agents Cli Workflow
This skill has Claude Code walk you through building an ADK agent step by step, from agreeing on the design through scaffolding, code and evals to deployment and monitoring.
What is Google Agents Cli Workflow?
Google Agents Cli Workflow is a skill for Claude Code and other coding agents that lays out the whole process of building agents with Google's agents-cli and the Agent Development Kit (ADK). It is the entry point of the agents-cli skills and splits the work into phases: understanding, studying recipes, scaffolding, building, evaluating, deploying, publishing and observing. For each phase it points Claude to the matching specialist skill.
Before any code is written, Claude talks with you about what the agent should do, which external APIs it needs, which limits apply and where it will run. It asks a single question per message and records the outcome as a spec that you approve. After that, Claude looks for matching recipes, for example for retrieval, memory across sessions or approval before risky actions, scaffolds the project and writes the code.
The skill suits developers who build agents in Python or Go and want to run them on Agent Runtime, Cloud Run or GKE. It requires the agents-cli command line tool in a version compatible with 1.15.0, which you install with 'uv tool install google-agents-cli'. You need uv for that. 'agents-cli info' shows you whether everything is in place.
Who it is for
Developers and platform teams who build AI agents with ADK in Python or Go and want to test, deploy and monitor them on Google Cloud.
Use cases
- A business unit wants an agent that searches its own documents, and you want to pin down the requirements before any code exists.
- You have built an ADK agent locally and want to check it with evals and then move it to Cloud Run or Agent Runtime.
- An agent is already in production, and you want to set up traces, logging and monitoring or narrow down bugs in the agent code.
What Google Agents Cli Workflow does
Phase plan with specialist skills
The skill maps every phase from understanding to observing to its own agents-cli skill, such as scaffolding, evals or deployment, and has Claude reload it before the phase starts.
Design dialogue before code
Claude asks one question per message, offers two or three architecture options for complex agents and only writes code once you have approved the spec.
Existing recipes first
For capabilities such as document search, sandboxed execution, memory across sessions, OAuth or voice agents, Claude first looks for a matching recipe and studies it before building anything itself.
Try and grade locally
'agents-cli playground' starts a web interface for testing, 'agents-cli lint' checks the code, and commands such as 'eval run' grade how your agent behaves.
Deployment with approval
Deployment to Agent Runtime, Cloud Run or GKE only runs after you agree. Optionally you set up CI/CD with GitHub Actions or Cloud Build and publish the agent to Gemini.
Say this to activate the skill
"Help me develop an agent with ADK.""Run my agent locally so I can try it out.""Check my agent with an evaluation.""Deploy my agent to Cloud Run." Install
npx skills add google/agents-cli --skill google-agents-cli-workflow
The command installs the original version from the author's repository (Vercel skills CLI, needs Node.js), add -g for all projects. We do not host a copy: updates come straight from the author. Need help with setup? How to install skills →
Figures with source and date
- GitHub stars
- ≈ 6,400
- Installs
- 578,974
- On skills.sh since
- 21 Apr 2026
- Last change in the repository
- 17 Nov 2026
- Licence
- Apache-2.0
As of 18 Nov 2026. Sources: GitHub, skills.sh. The figures change daily; we show the state of our last lookup.
Checked automatically
- Automated check on 2026-11-18 (revision 2fa159a): 7 files in the skill folder read by pattern search. It found no hidden characters, no instructions to override rules, no download-and-run and no access to credentials.
Independent checks on skills.sh, as of 18 Nov 2026
| Gen Agent Trust Hub | Pass |
|---|---|
| Socket | Pass |
| Snyk | Warn |
Good to know
- The skill needs agents-cli in a version compatible with 1.15.0 as well as uv. You update an older version through uv.
- The skill is set up to be always active and applies to any agent development, not only ADK projects.
- Claude only writes code after you approve the spec. Even for simple agents you get at least one clarifying question first.
- Deployment and publishing target Agent Runtime, Cloud Run, GKE and Gemini. The skill does not cover other cloud providers.
- One auditor on skills.sh reports a warning (Snyk). Read its notes before bringing the skill into projects with sensitive data.
- The check is an automated pattern search, not a security guarantee, and reads the skill folder only. The command always installs the author's current version; later versions have not been reviewed by us.
Frequently asked questions about Google Agents Cli Workflow
What is Google Agents Cli Workflow?
Google Agents Cli Workflow is the entry skill of Google's agents-cli. It gives Claude Code a fixed process for building agents with ADK, from the design dialogue through scaffolding, code and evals to deployment, publishing and monitoring.
How do I install Google Agents Cli Workflow?
You install Google Agents Cli Workflow with 'npx skills add google/agents-cli --skill google-agents-cli-workflow'. The skill needs the agents-cli tool, which you set up with 'uv tool install google-agents-cli'. 'uvx google-agents-cli setup' also installs the related skills.
Is Google Agents Cli Workflow free?
Google Agents Cli Workflow is released under the Apache-2.0 license and is free to use. The skill says nothing about the cost of the Google Cloud services you deploy your agent to.
Who is Google Agents Cli Workflow for?
Google Agents Cli Workflow is for teams that build agents with ADK in Python or Go and want to run them on Agent Runtime, Cloud Run or GKE. If you do not use Google Cloud, the design, build and evaluation phases are the most useful parts.
What data does Google Agents Cli Workflow send?
Google Agents Cli Workflow consists of instructions for Claude. Data leaves your machine through agents-cli commands, for example when you deploy, publish to Gemini or run evals in the cloud with 'eval submit'.
How does Google Agents Cli Workflow differ from the other agents-cli skills?
Google Agents Cli Workflow describes the overall process and hands off per phase. Details on code, deployment or publishing come from specialist skills such as google-agents-cli-adk-code, google-agents-cli-deploy and google-agents-cli-publish.
Share this skill
About this skill
Google Agents Cli Workflow comes from google and is an open-source project maintained by its author and community.
Author: google / github.com. Source: github.com/google/agents-cli/tree/main/skills/google-agents-cli-workflow. Licence: Apache-2.0.
Curated by Collective Brain, an AI and automation digital agency from Hamburg, together with WhiteFox Automations. More vetted skills in the skills catalog.
Working in Germany, Austria or Switzerland? We help teams roll out Claude, in English or German.
Want this skill tailored to your team?
Google Agents Cli Workflow is a strong starting point. A skill gets really powerful once it knows your workflows, templates and tone of voice. We build custom skills for teams, from analysis to rollout, and connect them to our work in AI consulting.
Request a custom skill →