# Google adds gcloud CLI, agent skill, and client libraries for its Developer Knowledge API

published: 2026-10-11T23:43:41.466Z
topic: infrastructure
source: https://developers.googleblog.com/supercharge-your-development-with-the-google-developer-knowledge-api-ecosystem/
url: https://aiagentnews.io/stories/google-adds-gcloud-cli-agent-skill-and-client-libraries-for-its-developer-knowle

Google has published a set of tools around its Developer Knowledge API, which provides official Google documentation to agents and developer tools. The toolset includes a gcloud CLI surface, an agent skill, API Explorer, and client libraries in several languages.

Google describes the Developer Knowledge API as its official programmatic source of documentation for Google Cloud, Firebase, Android and other products. The API offers semantic and keyword search, document chunking and grounded Q&A, and it returns fresh Markdown-formatted pages. Google says the API is indexed frequently to keep retrieval predictable and reduce the lag between documentation updates and agent access.

The gcloud CLI surface exposes answer-query, documents search-chunks and documents describe commands. It comes pre-installed in Cloud Shell and works with standard Google Cloud SDK installations on Linux, macOS and Windows. The agent skill, installed with npx skills add google/skills --skill retrieving-developer-knowledge, guides coding assistants such as Claude Code, Cursor and GitHub Copilot to use the Developer Knowledge MCP server or fall back to the REST API.

Client libraries are available for C#, Go, Java, Node.js and TypeScript, PHP, Python and Ruby. They provide access to AnswerQuery, SearchDocumentChunks, GetDocument and BatchGetDocuments, and BatchGetDocuments can retrieve up to 20 documents in one call. An interactive API Explorer is also available for testing requests and responses.

## Why it matters if you build agents

Agent builders can pull current Google Cloud and Firebase documentation through a CLI, an MCP server or a client library instead of scraping web pages or relying on model training data.

BatchGetDocuments accepts up to 20 documents per call, which can reduce request count when an agent needs several pages at once.

Teams should check authentication setup, since the libraries rely on Application Default Credentials or an API key depending on how the MCP server or client is configured.