shape

Overview of shape skill

What shape does

The shape skill is a planning-first UX/UI workflow for defining a feature before any code is written. Instead of jumping straight to layouts or components, it runs a structured discovery interview and turns the answers into a design brief. That makes it useful when your team knows what to build in broad terms but has not yet clarified user goals, constraints, states, edge cases, or implementation direction.

Who should install shape skill

The best fit for shape skill is product-minded builders, designers, frontend engineers, and AI-assisted teams who want better decisions before prototyping. It is especially relevant for shape for UI Design when a feature is still ambiguous: new flows, dashboards, forms, onboarding, settings, or any interface where “just generate a UI” would likely miss the real job users need done.

Why people choose shape over a generic prompt

A normal prompt often produces instant mockups or component ideas. shape instead slows the process down on purpose. Its differentiator is that it forces context gathering first, then creates a design brief you can hand off to implementation skills later. The repo is explicit about scope: this is design planning, not coding. That boundary is valuable if you want fewer shallow UI suggestions and more defendable product decisions.

How to Use shape skill

Install context and required prerequisite

Install the skill from the pbakaus/impeccable repository using your skills workflow, for example:
npx skills add pbakaus/impeccable --skill shape

The most important adoption detail is the prerequisite in SKILL.md: shape requires /impeccable first. The skill says you must invoke /impeccable, follow its Context Gathering Protocol, and if no design context exists yet, run /impeccable teach before using shape skill. If you skip that step, you are using the skill against its intended workflow.

What input shape needs to work well

The argument hint is [feature to shape], but a one-line feature name is usually not enough. Stronger shape usage starts with:

• the feature goal
• target user or role
• current workflow or pain point
• success criteria
• hard constraints like platform, permissions, compliance, or existing design system
• known edge cases or non-goals

A weak input is: Shape a notifications page.
A stronger input is: Shape a notifications center for account admins who miss urgent billing and security events. It must work on desktop first, reuse our existing table and filter patterns, and avoid adding real-time infrastructure in v1.

Practical workflow and files to read first

Read SKILL.md first and treat it as the operational contract. This repository snapshot only exposes that file, so most of the value is in following its sequence correctly:

1. Gather design context through /impeccable.
2. Use shape during planning, not implementation.
3. Let it conduct the discovery interview.
4. Turn the interview into a design brief.
5. Hand the brief to /impeccable craft, /impeccable, or another implementation workflow.

This matters because the skill is optimized for reducing guesswork before UI work begins, not for producing polished screens in one shot.

Prompt pattern that gets better outputs

To invoke shape skill effectively, ask for the interview and the final artifact explicitly. Example structure:

• feature to shape
• who the primary user is
• current problem
• must-have constraints
• what decisions you want the brief to settle

Example:
Use shape to plan a bulk-edit inventory feature for operations managers. Interview me first. Focus on user intent, error prevention, empty/loading/failure states, permissions, and what the v1 interaction model should be. Output a design brief I can hand to implementation.

This works better than “design a UI for X” because it gives the skill room to ask clarifying questions before locking in a direction.

shape skill FAQ

Is shape skill for coding or for planning?

shape skill is for planning. The repository is explicit: it does not write code. Its output is a design brief that guides later implementation. If you want immediate code, this is the wrong starting point; if you want better product and UI decisions first, it is a good fit.

When is shape better than ordinary prompting?

Use shape when the feature is underdefined, risky, user-facing, or likely to have tricky states and tradeoffs. A generic prompt may be faster for a throwaway mockup. shape is better when you need reasoning about workflow, user needs, constraints, and handoff quality.

Is shape suitable for beginners?

Yes, with one caveat: beginners still need enough product context to answer the discovery interview well. The skill gives structure, but it cannot invent your users, constraints, or success metrics. If you are new to UX planning, its interview-driven approach is actually helpful because it surfaces the questions you should have answered anyway.

When should I not use shape for UI Design?

Skip shape for UI Design when the problem is already fully specified, when you only need a tiny visual tweak, or when the task is purely technical and not about user interaction. It is also a poor fit if you refuse the prerequisite context-gathering step, because the skill depends on that foundation.

How to Improve shape skill

Give shape better planning inputs

The biggest quality lever is input quality. For shape, better inputs mean more than feature names. Include user type, task frequency, failure cost, existing patterns, business rules, and what should stay out of scope. The skill will still interview you, but richer starting context leads to a sharper brief and fewer generic recommendations.

Avoid the main failure mode: premature solutions

The repo’s strongest warning is not to make design decisions too early. A common misuse of shape skill is asking for screens, cards, tabs, or layouts before clarifying the user problem. If you notice the conversation jumping to interface patterns too soon, redirect it: ask for unmet user needs, task flow, states, tradeoffs, and constraints first.

Iterate on the design brief, not just the prompt

After the first pass, improve the brief by challenging unclear decisions:

• Which user goal is primary?
• What states are missing?
• What assumptions need validation?
• What is intentionally excluded from v1?
• What would make implementation ambiguous?

That kind of iteration improves shape usage more than repeatedly asking for “a better UI.”

Pair shape with downstream execution

The practical way to improve shape install value is to treat it as the first stage in a chain. Use shape to create the brief, then pass that brief into /impeccable craft or another implementation skill. The better your handoff artifact, the more likely later code or design generation will stay aligned with user needs instead of drifting into plausible-looking but weak UI.