Build What Agents Want: MCP, APIs, Skills & Plugins

That's exactly the right time. You'll leave with a framework for deciding which surfaces your product needs and in what order - and you'll have built each one, so the decision comes from experience rather than a blog post or YouTube.

YES! Some API experience is needed for this course.

Ideally, you've worked on an API Product already.

If you don't have API experience, you can always enroll in API Fundamentals for Product Managers, before taking this course.

As long as you have some familiarity with APIs and have worked on them in some capacity, that's fine.

But if you are completely unfamiliar with APIs, I'd recommend you enroll in API Fundamentals for Product Managers, prior to taking this course.

The course will be especially beneficial for those who already have a public API they are managing.

By the end: (1) two working, deployed MCP servers, (2) an MCP App ready for the ChatGPT or Claude Plugins Directory, and (3) an agent Skill for your product. You'll also have the frameworks and tools to redesign your APIs for agents and to build your roadmap by deciding which agent surfaces to prioritize.

Topics such as API Governance, Developer Experience, The API Lifecycle, API Standards/Guidelines and API Management require a lot more time.

Please see my API Product Mastery for Experienced PMs course to learn about these subjects.

Not necessarily, but it helps! All our hands-on lessons use no-code and the free version AI codegen tools like Zuplo, Replit, Fractal & MCPJam and you can use any LLM you have access to.

If you want to learn to code agents or train models, this isn't that course: we build with no-code and AI tools, and the focus is deciding what your product needs. If you can't set aside time to build, you'll miss most of the value, because every session ends with something built for your own product.

Yes! You'll learn where a CLI fits and when it beats an MCP server - it's part of the framework for deciding what your product needs, and cover what an agent-friendly CLI looks like.

Yes, for both Claude and Muse! I've an approved Claude Connector in the official directory and one submitted to Meta's Muse AI. They leverage the same MCP server and I'll teach you how to do that across Claude, Muse and also ChatGPT.

Yes! This course grew out of my experiences building MCP to be used as plugins in ChatGPT. I have been teaching this since early 2026 and have an approved ChatGPT Plugin published in the directory - check it out and install it!

Sessions will be recorded and will be made available on the Maven platform.

If you miss a session, please watch the recording and check out the slides.

You'll also have a chance to follow up with me after the course in a 1:1.

No. API Product Mastery for Experienced PMs covers API design, strategy, and governance across the full product lifecycle. This course starts where Mastery's agent-readiness section ends, going deep on MCP, evaluation, and MCP Apps. They complement each other. If you took Mastery, you'll get even more out of this course.

A workshop teaches you one surface. This teaches you all of them, on your product, so you can tell which ones you actually need - that comparison is the whole point and no single workshop can give it to you. You also get a cohort, a 1:1 with me afterward, and a group session a month later. Workshop alumni get an alumni discount - ping me.

This will be five live 3-hour sessions over Zoom. The course is delivered over two weeks, and you can optionally book a 1:1 meeting with the instructor. After one month, the group will meet again to show what they have built and get feedback.

Many students get reimbursed through L&D or professional development budgets. Reach out if you need help making the case to your manager, or use the Maven expense template.