google/skills

This repository is Google's own official collection of Agent Skills for Google Cloud and related Google products. An Agent Skill is a packaged bundle of instructions and reference material that an AI coding assistant can load to become knowledgeable about a specific tool or task, and this repo gathers dozens of them covering Google Cloud in particular. You install it with a single command, and that command lets you pick which specific skills from the collection you actually want, rather than pulling in everything at once. Once installed, your AI assistant can use these skills whenever a relevant task comes up, without you having to re-explain how a given Google Cloud service works each time. The skills are organized into groups that mirror how people actually work with Google Cloud. There are getting started skills for authenticating and onboarding to Google Cloud, multi-product solution skills that walk through architecture patterns like building a data lakehouse or deploying AI agents, and a large AI and machine learning section covering things like the Gemini API, model tuning, and the agent platform's various management tools. A big infrastructure section focuses heavily on Google Kubernetes Engine, with skills for cluster creation, networking, autoscaling, cost optimization, and troubleshooting specific failure modes. Further sections cover databases and analytics tools such as BigQuery, AlloyDB, Bigtable, and Spanner, developer tools including the gcloud command line interface, and management tools for logging, monitoring, and alerting. Because it is maintained directly by Google and explicitly noted as under active development, the list of available skills is expected to keep growing over time. The project is written primarily in Python and is aimed at anyone using an AI coding assistant to work with Google Cloud, whether they are setting up infrastructure for the first time or troubleshooting an existing production system.