Vercel has introduced Agent Plugins 1.0.0, a vendor-neutral format for packaging reusable AI-agent extensions so compatible clients can discover and load them in a consistent way.

The first version focuses on two portable component types: Agent Skills and Model Context Protocol servers. The format defines a plugin directory with a plugin.json manifest at the root, optional Skills under skills/, and MCP configuration in mcp.json. Vercel says the manifest needs only the targeted specification version and plugin name at minimum, while component validation happens separately so one invalid piece does not disable unrelated parts.

The launch is narrower than a full agent runtime. Distribution, installation, permissions, policy, user experience, and client-specific behavior remain under each client team's control. The specification also includes a namespaced extension mechanism so clients can keep proprietary or experimental capabilities outside the shared portable contract.

That boundary is important for developers who already maintain the same Skill or MCP server across several agent products. A small common package shape could reduce repackaging work without forcing clients to standardize how they expose tools to users or models.

Vercel says it initiated the proposal with input from representatives at AWS, Anysphere, GitHub, Microsoft, OpenAI, and Vercel. The project site says the initial Technical Steering Committee includes Core Maintainers from Amazon, Cursor, Microsoft, OpenAI, and Vercel. At launch, Vercel lists support across ChatGPT and Codex, Cursor, GitHub Copilot, Kiro, and VS Code.