That shortens the argument. An agent can draft a spec from the repo and place a code layer on the canvas, so design and engineering are looking at the same object. Calling generate_image spends AI credits. Using the agent on a spec spends credits too. Hover the receipt if you want the number.
The component that ships still has to match the package: same props, same tokens, same states. MCP does not merge the pull request. A person does, against the package the app installs.
Until a code source is attached
Until a code source is attached to your library, the agent is guessing your components. The prop list in the guidelines file gives that guess a document. If the system and the product sit in different GitHub organizations, connect both on an Organization or Enterprise plan before you ask the agent to use the system.
I led Aurora at Nord, used by dozens of teams and millions of users. The painful weeks were rarely a missing component. A prop meant one thing in Figma and another in React. An agent on the canvas has to pick a source. The review is where you catch the pick.
I checked the Figma release notes and the Config India 2026 help article on 18 Oct 2026. Plans, betas, and the 4 Nov credit window come from those pages. If your admin console disagrees, trust the console.