--- title: Behind the scenes: How we build, test, and scale Google... description: Google Agent Skills is an open-source repository that encodes Google Cloud domain knowledge into structured instructions for AI coding agents. The post details... canonical: https://daily.dev/posts/behind-the-scenes-how-we-build-test-and-scale-google-agent-skills-xz3oqrnps twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Behind the scenes: How we build, test, and scale Google Agent Skills | daily.dev og:description: Google Agent Skills is an open-source repository that encodes Google Cloud domain knowledge into structured instructions for AI coding agents. The post details... og:url: https://daily.dev/posts/behind-the-scenes-how-we-build-test-and-scale-google-agent-skills-xz3oqrnps og:image: https://api.daily.dev/og/posts/xz3OQRnPS.png og:image:alt: Behind the scenes: How we build, test, and scale Google Agent Skills og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Behind the scenes: How we build, test, and scale Google Agent Skills **[Google Cloud](https://daily.dev/sources/gcp)** · 5 min read · 5 upvotes · 1 comments ## Summary Google Agent Skills is an open-source repository that encodes Google Cloud domain knowledge into structured instructions for AI coding agents. The post details the engineering process behind it: a standardized skill layout, automated CI/CD checks (linters, link checkers, AI-assisted validation), and continuous evaluation pipelines that run on submit and weekly to catch regressions. Skills are treated as living products with designated owners, not one-off documents. The team also built internal agentic authoring tools using ADK and a parallel internal initiative called DevRel Skills for team workflows. The project launched at Google Cloud Next 2026 and received over 15,000 GitHub stars. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is the recommended architecture for building an agent skill that needs to call external tools? Prefer referencing remote Model Context Protocol (MCP) tools whenever possible, and only fall back to CLI or API calls when necessary. Remote MCP servers suit agentic workloads because they provide tools while also offering built-in authentication and IAM governance, making them the preferred choice for Google Agent Skills. _daily.dev surfaces architectural guidance like MCP-first design for teams building agent tooling._ ### How does Google evaluate whether an agent skill actually improves AI agent performance before publishing it? Skill authors must supply evaluation prompt suites with scoring rubrics, and every new skill is evaluated internally comparing agent performance with and without the skill. Evaluations measure two dimensions: accuracy (response quality and task completion rate) and efficiency (tokens consumed and completion time), run across multiple agent frameworks and repeated weekly to catch regressions. _teams weighing how to validate AI agent instructions can track evaluation practices like this on daily.dev._ ### How many GitHub stars did Google's Agent Skills repository get after launch? The Google Agent Skills repository exceeded expectations by gaining over 15,000 GitHub stars after its initial launch, part of a fast-paced effort that began ahead of Google Cloud Next 2026 led by Developer Advocates and Technical Writers packaging Google Cloud domain knowledge into agent-readable instructions. _developers gauging traction of new agent tooling projects can follow milestones like this on daily.dev._ ## Community discussion Top comments from developers on daily.dev. **@agustinbarrientos** · 0 upvotes > The weekly cross-framework evals sound excellent; where do they usually find the biggest differences for the same skill? ## Similar posts on daily.dev - [Agent Skills explained: Replacing AGENTS.md with reusable AI knowledge](https://daily.dev/posts/agent-skills-explained-replacing-agents-md-with-reusable-ai-knowledge-4lgrqdva9) · SwiftLee · 1 upvotes · 0 comments - [Developer’s Guide to Building ADK Agents with Skills](https://daily.dev/posts/developer-s-guide-to-building-adk-agents-with-skills-ofskaw4tw) · Google Developers · 1 upvotes · 0 comments - [Agent Skills](https://daily.dev/posts/agent-skills-oiotdzetc) · Hacker News · 0 upvotes · 0 comments - [Google Cloud launches GEAR program to broaden AI agent development skills](https://daily.dev/posts/google-cloud-launches-gear-program-to-broaden-ai-agent-development-skills-9c4sjpzli) · InfoWorld · 0 upvotes · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents), [#cicd](https://daily.dev/tags/cicd), [#gcp](https://daily.dev/tags/gcp), [#mcp](https://daily.dev/tags/mcp) [View this post on daily.dev](https://daily.dev/posts/behind-the-scenes-how-we-build-test-and-scale-google-agent-skills-xz3oqrnps) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Behind the scenes: How we build, test, and scale Google Agent Skills","url":"https://daily.dev/posts/behind-the-scenes-how-we-build-test-and-scale-google-agent-skills-xz3oqrnps","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/behind-the-scenes-how-we-build-test-and-scale-google-agent-skills-xz3oqrnps"},"datePublished":"2026-08-03T14:24:15.511Z","dateModified":"2026-09-14T06:41:16.236Z","description":"Google Agent Skills is an open-source repository that encodes Google Cloud domain knowledge into structured instructions for AI coding agents. The post details...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/aea4e0aa1da83ed7b95caebbc1aa13cf?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/aea4e0aa1da83ed7b95caebbc1aa13cf?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Google Cloud","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Google Cloud","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/gcp","url":"https://daily.dev/sources/gcp"},"commentCount":1,"discussionUrl":"https://daily.dev/posts/behind-the-scenes-how-we-build-test-and-scale-google-agent-skills-xz3oqrnps","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":5},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":1}],"keywords":"llm,ai-agents,cicd,gcp,mcp","timeRequired":"PT5M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Google Cloud","item":"https://daily.dev/sources/gcp"},{"@type":"ListItem","position":3,"name":"Behind the scenes: How we build, test, and scale Google Agent Skills"}]} {"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/behind-the-scenes-how-we-build-test-and-scale-google-agent-skills-xz3oqrnps","comment":[{"@type":"Comment","text":"The weekly cross-framework evals sound excellent; where do they usually find the biggest differences for the same skill?","datePublished":"2026-08-03T17:39:56.385Z","url":"https://daily.dev/posts/xz3OQRnPS#c-EJCJHVbd6","author":{"@type":"Person","name":"Agustin Barrientos","url":"https://daily.dev/agustinbarrientos","image":"https://media.daily.dev/image/upload/s--5ayxQnqn--/f_auto/v1788281802/avatars/avatar_wQYYVe5Tbj0NJ7C7qPoa8?_a=BAMAMicg0"}}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/behind-the-scenes-how-we-build-test-and-scale-google-agent-skills-xz3oqrnps#faq","mainEntity":[{"@type":"Question","name":"What is the recommended architecture for building an agent skill that needs to call external tools?","acceptedAnswer":{"@type":"Answer","text":"Prefer referencing remote Model Context Protocol (MCP) tools whenever possible, and only fall back to CLI or API calls when necessary. Remote MCP servers suit agentic workloads because they provide tools while also offering built-in authentication and IAM governance, making them the preferred choice for Google Agent Skills. daily.dev surfaces architectural guidance like MCP-first design for teams building agent tooling."}},{"@type":"Question","name":"How does Google evaluate whether an agent skill actually improves AI agent performance before publishing it?","acceptedAnswer":{"@type":"Answer","text":"Skill authors must supply evaluation prompt suites with scoring rubrics, and every new skill is evaluated internally comparing agent performance with and without the skill. Evaluations measure two dimensions: accuracy (response quality and task completion rate) and efficiency (tokens consumed and completion time), run across multiple agent frameworks and repeated weekly to catch regressions. teams weighing how to validate AI agent instructions can track evaluation practices like this on daily.dev."}},{"@type":"Question","name":"How many GitHub stars did Google's Agent Skills repository get after launch?","acceptedAnswer":{"@type":"Answer","text":"The Google Agent Skills repository exceeded expectations by gaining over 15,000 GitHub stars after its initial launch, part of a fast-paced effort that began ahead of Google Cloud Next 2026 led by Developer Advocates and Technical Writers packaging Google Cloud domain knowledge into agent-readable instructions. developers gauging traction of new agent tooling projects can follow milestones like this on daily.dev."}}]} ```