Introducing Agent Plugins

YouTube AI Channels Tools

Summary

AWS, Cursor, GitHub, Microsoft, OpenAI, and Vercel jointly launched Agent Plugins—an open, vendor-neutral packaging format for agent extensions, designed to unify the packaging and discovery of skills and MCP servers, promoting reuse across products.

No content available
Original Article
View Cached Full Text

Cached at: 08/06/26, 04:46 PM

TL;DR: AWS, Cursor, GitHub, Microsoft, OpenAI, and Vercel jointly launched Agent Plugins — an open, vendor-neutral packaging format for agent extensions that lets Skills and MCP servers be reused across products. ## Background: From Skills and MCP to Plugins Skills and MCP servers are making agents more powerful, and Plugins give developers a way to package these capabilities for sharing and reuse. But the problem is: today, every agent product expects a different manifest, folder structure, and setup process. This fragmentation increases the burden on plugin developers and limits the potential for cross-platform reuse. ## What Agent Plugins Is Agent Plugins is an open, vendor-neutral way to package agent extensions. It was created collaboratively by contributors from AWS, Cursor, GitHub, Microsoft, OpenAI, and Vercel. Its goals are simple: - Provide a unified format for people building plugins - Provide a predictable discovery mechanism for agent tools In its simplest form, an Agent Plugin is a folder with a manifest file named `plugin.json` at its root. ## This First Release Focuses on Two Capabilities This initial release focuses on two things developers are already using: 1. **Agent Skills**: for reusable instructions and workflows. 2. **MCP servers**: to connect agents to tools and data. Developers can place these resources in standard locations so that the same package is more likely to be supported across different products. ## Value and Benefits For plugin authors, the benefits are: - Less platform-specific glue code - A unified versioned format to build against For agent products: - Support skills, MCP, or both - Extend the format without changing its portable core ## Standardization Scope: Packaging and Discovery, Not Runtime Note that the Agent Plugins specification standardizes **packaging and discovery**, not marketplaces, permissions, or runtimes. In other words, it only addresses "how to organize, describe, and find plugins," leaving marketplace mechanisms, permission models, and execution environments to each product. ## Open Ecosystem and Next Steps Agent Plugins is open and can be built with right away. The recommended path is: 1. Package your capabilities 2. Add support in your product 3. Check it out for yourself Start at agentplugins.org (https://agentplugins.org). ## Source Introducing Agent Plugins - YouTube (https://www.youtube.com/watch?v=UaeWJK_vv-Y)

Similar Articles

Agent Plugins (4 minute read)

TLDR AI

Vercel announces Agent Plugins 1.0.0, an open, vendor-neutral standard for packaging Agent Skills and MCP servers into distributable plugins, giving AI agents a common format for discovery and loading.

@mattpocockuk: Oh sweet Expect a mattpocock/skills plugin ASAP

X AI KOLs Timeline

OpenAI announces Agent Plugins, an open standard developed with AWS, Cursor, GitHub, Code, and Vercel to package Agent Skills and support MCP server configurations, enabling plugins to work across compatible agent clients. Matt Pocock reacts enthusiastically and expects a skills plugin soon.

@wsl8297: Microsoft open-sourced a more convenient VS Code extension: AI Toolkit, which streamlines Agent application development from environment setup to evaluation deployment in one go. A single interface is enough: model selection, Playground trial, Agent building, batch comparison, metric evaluation, all integrated. Gi…

X AI KOLs Timeline

Microsoft open-sourced the VS Code extension AI Toolkit, which integrates model selection, Playground, Agent building, batch testing, and evaluation into one interface, simplifying AI Agent development process.