# google/skills for Claude Code: what it does and how to install it > google/skills is Google's own collection of Claude Code skills for its cloud and developer platforms. It indexes 126 components, 125 skills and 1 MCP server, covering Google Kubernetes Engine, Agent Platform, Genkit, major databases, logging, monitoring, IAM, and the Google Ads, Data Manager, and Analytics APIs. Use it when working directly with any of these Google products. Canonical: https://www.flowy.sh/plugins/google-skills/guide Published: 2026-10-29 Updated: 2026-10-29 Author: Flowy ## What is google/skills, and who is it for? `google/skills` is Google's own repository of Agent Skills, covering Google Cloud infrastructure, Google's AI and agent tooling, and a handful of Google's advertising and analytics APIs. The repository describes itself as under active development. Flowy indexes it as one listing of 126 components: 125 skills and 1 MCP server, filed under the development category. It behaves less like one plugin with a single job and more like a shelf of reference manuals: install the pieces for the Google product you're using and skip the rest. ## Which google/skills cover Google Kubernetes Engine? Thirty-one skills carry a gke name, more than a quarter of everything in this plugin, covering nearly every stage of running a cluster. `gke-basics` and `gke-cluster-creation` provision a cluster and choose between Autopilot and Standard; `gke-networking`, `gke-storage`, and `gke-service-networking` handle core infrastructure; `gke-cluster-autoscaler`, `gke-workload-scaling`, and `gke-compute-classes` cover scaling and machine selection; `gke-cost-analysis` and `gke-cost-optimization` answer cost questions against billing exports. `gke-platform-security` and `gke-workload-security` harden the cluster and its workloads, `gke-backup-dr` and `gke-upgrades` handle lifecycle events, and `gke-workload-troubleshooting` diagnoses pods that fail to start or crash on repeat. Five more skills specialize in GPU and TPU failure modes for AI training jobs, such as `gke-ai-troubleshooting-jobset-interruption` for restart loops and spot VM preemptions, and `gke-productionize` runs a full production readiness review. ## Does it help with building and running AI agents? Yes, through two families worth telling apart. Thirteen skills manage the lifecycle of a model or agent on Google's Agent Platform, the product formerly called Vertex AI: `agent-platform-deploy` and `agent-platform-endpoint-management` put a model into production, `agent-platform-model-registry`, `agent-platform-prompt-management`, and `agent-platform-tuning` manage its inputs, and `agent-platform-eval-flywheel` measures and improves quality once it's running. Four more skills call Gemini's APIs directly: `gemini-api`, `gemini-agents-api`, `gemini-interactions-api`, and `gemini-live-api` for a websocket streaming client. Separately, four skills, `genkit-dart`, `genkit-go`, `genkit-js`, and `genkit-python`, cover Genkit, Google's open source app framework for AI features, independent of Agent Platform. `google-agents-cli-onboarding` is the entry point for Google's agent development CLI, and ten more skills produce architecture guidance for a specific agent pattern, from `google-cloud-solution-multi-agent-security` to `google-cloud-solution-build-deploy-agents`. ## What about databases and data pipelines? Each major Google Cloud database gets its own basics skill: `alloydb-basics`, `bigquery-basics`, `bigtable-basics`, `cloud-sql-basics`, and `spanner-basics`, plus `cloud-databases-onboarding` to help pick between them. `bigquery-ai-ml` and `bigquery-bigframes` add BigQuery's built-in machine learning SQL and its pandas-style DataFrame API. Two skills trace how a change to a BigQuery table ripples downstream: `datalineage-bigquery-asset-impact-analysis` for blast radius analysis and `datalineage-summary` for a plain-language lineage report. Three more skills cover Airflow on Cloud Composer end to end: authoring (`managed-airflow-dag-authoring`), troubleshooting (`managed-airflow-dag-troubleshooting`), and version migration (`managed-airflow-migrations`). ## Does google/skills cover observability, security, and the Well-Architected Framework? Logging and monitoring split into seven skills. Three handle logging, `cloud-logging-configuration-basics`, `cloud-logging-cross-project-configuration`, and `cloud-logging-query-generation`, and four handle monitoring: `cloud-monitoring-promql-query`, `cloud-monitoring-list-time-series-request`, `cloud-monitoring-metric-selection`, and `cloud-monitoring-chart-generation`. `google-cloud-scc-query` pulls findings from Security Command Center, and `iam-helper-for-policy-simulator` replays real historical access logs against a proposed IAM policy before it ships, so a change doesn't quietly break something already running. Six skills map directly to the six pillars of the Google Cloud Well-Architected Framework, from `google-cloud-waf-cost-optimization` to `google-cloud-waf-sustainability`. ## Does google/skills cover advertising and analytics too? A separate slice of this plugin has nothing to do with cloud infrastructure. Five skills cover the Google Mobile Ads SDK: `google-mobile-ads-banner`, `google-mobile-ads-interstitial`, and `google-mobile-ads-rewarded` for the three ad formats, plus `google-mobile-ads-get-started` and `google-mobile-ads-android-migrate-to-next-gen` for first integration and migrating off the older SDK. Two more, `ima-dai-sdk` and `ima-sdk-client-side`, cover video ad insertion. Three skills serve advertisers directly through the Google Ads API: `google-ads-api-quickstart` for credential setup, `google-ads-api-account-diagnostics` for diagnosing performance drops, and `google-ads-api-mcp-setup`, which walks through connecting an AI assistant, Gemini, Claude Code, or Cursor, to a live Google Ads account. Three more skills, `data-manager-api-setup`, `data-manager-api-audience-ingestion`, and `data-manager-api-event-ingestion`, handle audience and conversion data through the Data Manager API, and `google-analytics-admin-api-basics` plus `google-analytics-data-api-basics` cover Analytics account configuration and reporting. ## What does the developer-knowledge MCP server do? The plugin's single MCP server, `developer-knowledge`, pairs with the `retrieving-developer-knowledge` skill to search Google's own developer documentation, spanning Cloud, AI and Gemini, Android, Chrome, Web, Flutter, and Firebase, from inside a session instead of a browser tab. Before any of that, `gcloud` adds validation and guardrails to any command it helps generate, and three onboarding skills, `google-cloud-recipe-auth`, `google-cloud-recipe-onboarding`, and `google-cloud-recipe-foundation-builder`, cover authentication, first project setup, and deploying a full landing zone foundation for a new Google Cloud organization. ## How do I install google/skills in Claude Code? Install from [the install card on the listing page](https://www.flowy.sh/listings/google-skills#install). Flowy indexes what `google/skills` ships; it does not own, bundle, or host the repository itself. With 125 skills spread this wide, most people will only need a handful installed at once. Search or browse by the product you're actually working with rather than pulling in all of it. Sources: [google/skills on GitHub](https://github.com/google/skills), [google/skills README](https://github.com/google/skills#readme) ## Common questions ### What does the google/skills plugin cover in Claude Code? google/skills is Google's own repository of Claude Code skills for its cloud and developer platforms. It indexes 126 components, 125 skills and 1 MCP server, spanning Google Kubernetes Engine, Agent Platform, Genkit, Google Cloud databases, logging and monitoring, IAM, and the Google Ads, Data Manager, and Analytics APIs. ### Does google/skills help with Google Kubernetes Engine (GKE)? Yes. Skills carrying a gke name are the largest group in the plugin, thirty-one in total, covering cluster creation, networking, storage, autoscaling, cost analysis, security hardening, backups, upgrades, and workload troubleshooting. A further five skills focus specifically on GPU and TPU failure modes for AI training workloads on GKE. ### What is the difference between the Agent Platform skills and the Genkit skills? The agent-platform skills manage models and agents running on Google's Agent Platform, the product formerly called Vertex AI: deployment, endpoints, tuning, prompts, and evaluation. The four Genkit skills, `genkit-dart`, `genkit-go`, `genkit-js`, and `genkit-python`, cover Genkit, Google's separate open source app framework for building AI features, which does not require Agent Platform at all. ### Can I use google/skills to connect an AI assistant to Google Ads? Yes. `google-ads-api-mcp-setup` walks through installing Google's official open source Google Ads MCP Server so an assistant such as Gemini, Claude Code, or Cursor can query campaigns or pull account data directly. It sits alongside two other Google Ads API skills covering account diagnostics and a general quickstart. ### What does the developer-knowledge MCP server do in this plugin? It is the plugin's only MCP server, and it pairs with the `retrieving-developer-knowledge` skill to search and synthesize Google's own developer documentation, across Cloud, AI and Gemini, Android, Chrome, Web, Flutter, and Firebase, from inside a Claude Code session instead of a separate browser tab. ## Related - [The listing page](https://www.flowy.sh/listings/google-skills): Install command, components and repository - [Alternatives to this plugin](https://www.flowy.sh/alternatives/google-skills): Peers in its category, ranked by stars - [The Claude Code plugin marketplace](https://www.flowy.sh/marketplace/plugins): How plugins install, and every plugin in the index - [superpowers for Claude Code: what it does and how to install it](https://www.flowy.sh/plugins/superpowers-flow/guide): another plugin guide - [mattpocock/skills for Claude Code: install and use](https://www.flowy.sh/plugins/matt-pocock-flow/guide): another plugin guide