Google has joined the core maintainer group behind Agent Plugins, an open specification for packaging AI-agent skills and Model Context Protocol servers for use across compatible clients, placing it alongside Amazon and Microsoft in the project's technical governance. Google announced its participation Aug. 6, with Kevin Hou, a senior staff engineer at Google DeepMind, representing the company as a core maintainer. The Google Developers Blog announcement said Agent Plugins 1.0.0 was published by a technical steering committee with representatives from Amazon, Cursor, Microsoft, OpenAI and Vercel. The specification addresses client-specific packaging differences, described by Google as "The core problem isn't the components. It's the manifest." The result is a vendor-neutral package format rather than another agent runtime. The 1.0.0 specification was published July 24, while Google's Aug. 6 announcement marked its addition to the maintainer group and a broader public rollout. An Agent Plugin is a self-contained directory with a required plugin.json manifest at its root, potentially containing Agent Skills under skills/, MCP server configuration in mcp.json, and client-specific files under reverse-domain namespaces. The minimum manifest contains a canonical schema identifier and a plugin name, with optional fields for version, description, author, homepage, repository, license and keywords. The schema is closed, with client-specific data under an extensions object. The specification defines failure boundaries, where a missing optional component location is not an error, and invalid skills or