Spec Scope Seed Ship
Spec Scope Seed Ship is a four part prompt pattern that turns vague AI requests into review ready code. You open with the exact framework, component library, design tokens, grid, type scale and destination file. Then you declare what is out of bounds. Next you give a concrete visual example to copy. You close by listing the conditions the model must meet before it can claim done. The pattern was battle tested across v0, Cursor, and Lovable builds in 2026.
This is not a wish list. It is a design brief for a junior developer who happens to be an AI. Most designers start with adjectives and vibes. They get generic output that requires ten rounds of fixes. The four part structure collapses those rounds into one or two.
Spec Scope Seed Ship is not four separate prompts. It is one single message that contains all four sections in order. Treating them as disconnected steps breaks the context window and produces inconsistent results. The power lives in the combination.
Common confusion comes from thinking more tokens equal better output. The opposite holds. A tight, well structured prompt that opens with Spec, sets Scope, offers Seed, and ends with Ship beats a rambling three paragraph description every single time. The model stops hallucinating when the rules are explicit.
Concrete example. A designer building a hero section writes the following. Spec: Next.js 15 App Router, shadcn, Tailwind 4, tokens in globals.css, ship to app/(marketing)/hero/page.tsx. Scope: Do not modify any file outside this route. Do not add new dependencies. Use the existing Button component. Seed: Match the visual style, shadow scale, and spacing rhythm of the HeroSection component in our main repo. Ship: Done when the headline uses the correct weight hierarchy, the CTA button is focusable by keyboard, the mobile layout stacks without overlap, and there are zero TypeScript errors.
That single prompt generated production ready code in v0 that survived the first engineer review with only minor tweaks. The same pattern applied to Bolt produced a full signup flow with matching visual language. The designer reused the exact prompt skeleton across seven different tools and maintained consistency without writing CSS.
Use Spec Scope Seed Ship when you want one shot generations that pass diff review. It earns its keep on repeatable components, marketing pages, and feature extensions where consistency matters more than surprise. The pattern forces you to think like a systems designer before you touch the keyboard.
Do not use this pattern when you are exploring completely new territory with no existing design system or codebase patterns. In true zero to one discovery phases a more open ended conversation sometimes surfaces better directions. The tradeoff is speed and consistency versus open exploration. Know which phase you are in.
Teams that adopted this pattern in 2026 reported cutting iteration cycles by 70 percent. They stopped treating the AI like an oracle and started treating it like a very fast junior who needed clear tickets. The output quality jumped because the prompts began to resemble the same specs they once gave human collaborators.
The pattern also scales. Once you have a working Spec Scope Seed Ship template you can create variants for different tools. v0 needs heavier visual seeds. Cursor needs stricter scope because it edits real repos. Lovable needs clear database and auth requirements listed in the Ship section. Adjust the weight of each part but never drop one.
Mastering Spec Scope Seed Ship is the difference between vibe coding as a party trick and vibe coding as a repeatable shipping process. It forces the discipline the meme version conveniently ignores.
Your prompts stop reading like mood boards and start reading like tickets a senior engineer would respect. That single change is what lets solo designers ship code that survives code review.
Read the full guide
Related terms
Keep exploring
Vibe Coding
Vibe coding is designers using natural language to describe the exact feeling behavior and references they want then watching AI generate real React Tailwind or shadcn code instead of static Figma frames.
Prompt Engineering
The practice of writing instructions that produce consistent, usable output from a language model. Functionally identical to writing a good creative brief.
Design Tokens
The atomic design values (colors, spacing, typography, shadows, motion) stored as platform-agnostic variables that every component in a design system references.
Component Library
A collection of reusable UI elements (buttons, inputs, cards, modals) built from design tokens and documented with usage guidelines. One layer of a design system, not the whole thing.