Guide · Agent skill
How to use the Build and Deploy Google Cloud AI Agents agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
This page explains how to install and use the google-cloud-solution-build-deploy-agents skill, a set of instructions that lets an AI agent design, plan and document an AI agent or multi-agent system on Google Cloud. It is written for readers who have not used agent skills before. By the end you will know what the skill does, how to add it to your agent, how to phrase a request so it triggers, and what to check before running any cloud commands.
Build and Deploy Google Cloud AI Agentsgoogle/skills · Apache-2.0 Download the skill Quick start
- 1Download the ZIP. Download the skill ZIP from the skill's page on this site.
- 2Unzip the folder. Unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder.
- 3Or upload it in the Claude apps. Go to Settings, then Capabilities, then Skills, and upload the skill.
- 4Describe your agent system. Write a request that names the workload and asks for an architecture, diagram or deployment plan.
- 5Answer the clarifying questions. Reply to requirements, dependency and ambiguity questions so the agent can recommend a pattern.
- 6Confirm each phase. Approve the technical decomposition, architecture and implementation plan before the agent moves on.
- 7Review the validation plan. Check the validation checklist and run only the commands you have reviewed.
What is this skill and what does it make the agent do?
An agent skill is a folder of instructions and templates that an AI agent reads when a task matches the skill's description. This one covers designing and implementing agentic systems on Google Cloud. Once installed, the agent follows a four-phase process rather than answering freely.
The four phases are requirements discovery and analysis, solution design, implementation plan, and solution validation. The agent asks clarifying questions about functional and non-functional requirements, existing systems and dependencies, and resolves any contradictions before it recommends an architecture. Each phase waits for your confirmation before moving on.
- Phase 1: gather requirements and recommend a single-agent or multi-agent pattern.
- Phase 2: map components to Google Cloud products and produce a Mermaid architecture diagram.
- Phase 3: generate deployment instructions.
- Phase 4: produce a validation plan for the deployed system.
How do you install the skill?
You can install the skill in one of two ways. Download the ZIP file from the skill's page on this site and unzip it. For a personal setup, place the unzipped folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder.
If you use the Claude apps instead, open Settings, go to Capabilities, and upload the skill under Skills.
- Personal install: unzip into ~/.claude/skills/.
- Project install: unzip into the project's .claude/skills/ folder.
- Claude apps: Settings → Capabilities → Skills, then upload.
What do you need before you start?
The skill itself only generates design documents and instructions. For any deployment steps you need a Google Cloud project and an account with access to it, the gcloud CLI, and, if the agent runs those commands, the Google Cloud REST API or MCP tools.
The skill uses the network, so the agent may need network access to look at Google Cloud guidance. It does not run scripts.
- A Google Cloud project and account for deployment steps.
- The gcloud CLI.
- Google Cloud REST API or MCP tools, if the agent will run those commands.
How do you trigger the skill with a natural request?
Describe the agent system you want to build and say that you want an architecture, diagram or deployment plan. The skill is meant for requests about designing or implementing agentic systems on Google Cloud. If your prompt already contains enough requirements and asks for an architecture proposal or diagram, the agent will skip the long question list and go straight to recommending a pattern.
If your request is vague, expect questions. That is intended: the skill asks instead of guessing.
- Design a multi-agent system on Google Cloud for [your workload], and produce an architecture diagram.
- Help me plan a single AI agent on Google Cloud that answers questions from our internal documentation, with a deployment plan.
- I want to build an agent on Google Cloud. Ask me what you need to know first.
What gets produced and what it does not do?
The skill produces a written solution architecture with a Mermaid diagram, an implementation plan with deployment instructions, and a validation plan with checks. The validation template, for example, includes a Terraform dry-run step (terraform plan) and a curl connectivity check as illustrations.
The skill does not deploy to a live project on its own. The agent must still run any gcloud commands it proposes, and those commands can change real cloud resources. Review every command before you approve it.
- Output: solution-architecture.md plus implementation and validation documents.
- Not included: automatic deployment to a live Google Cloud project.
What safety points should you know?
The skill's instructions tell the agent to read its attribution reference before running any gcloud, REST API or MCP command. Treat any command as a proposal until you have read it and agreed to run it. The skill runs no scripts, but it does use the network.
For general Google Cloud architecture that is not about AI agents, the SKILL.md points to a different skill, google-cloud-solution-architecture.
- Read every gcloud command before running it.
- Check which project and account are active before any deployment step.
- Use the validation plan's dry-run checks before applying changes.
What people use it for
- Planning a single agent that answers questions from internal product documentation on Google Cloud.
- Designing a multi-agent workflow where specialised agents collaborate on a business process.
- Producing a written architecture and Mermaid diagram to share with a team before building.
- Creating a deployment checklist and validation plan to review with your cloud operations staff.
Frequently asked questions
How do I install the google-cloud-solution-build-deploy-agents skill?
Download the ZIP from the skill's page, then unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder. In the Claude apps, you can instead upload it under Settings, then Capabilities, then Skills.
What do I need to use this skill?
The skill itself only generates documents. For deployment steps you need a Google Cloud project and account, the gcloud CLI, and Google Cloud REST API or MCP tools if the agent will run those commands. It uses the network and runs no scripts.
Does this skill deploy my agent to Google Cloud automatically?
No. The files describe generating instructions and templates. The agent must still run any gcloud commands it proposes, and those can change real cloud resources, so review them first.
Is this skill free and what is its licence?
The skill is listed under the Apache-2.0 licence. Check the licence file in the downloaded package for the full terms.
What is the difference between this skill and google-cloud-solution-architecture?
This skill is for designing AI agent or multi-agent systems on Google Cloud. The SKILL.md says general Google Cloud solution architecture that is not about agents should use google-cloud-solution-architecture instead, and narrow tasks about one product without agent context are also out of scope.
Why does the agent keep asking me questions?
The skill requires that ambiguities and contradictions be resolved before any architecture or product recommendation is made. If your request lacks requirements, expect questions about functional and non-functional needs, data sources and dependencies. Giving more detail in your first message usually shortens this.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
This skill suits engineers and teams who want a structured, confirmed design, product mapping and validation checklist before building an AI agent on Google Cloud. It is not suited to general cloud architecture, single-product tasks, or anyone expecting automatic deployment to a live project. Review every gcloud command the agent proposes before running it.
Download the skill