design-lab_skill

This skill helps you explore UI design options by generating multiple variations, collecting feedback, and producing actionable implementation plans.
  • TypeScript

520

GitHub Stars

2

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill 0xdesign/design-plugin --skill design-lab

  • DESIGN_PRINCIPLES.md53.2 KB
  • SKILL.md29.5 KB

Overview

This skill runs a temporary design lab to explore UI options, collect feedback, and produce actionable implementation plans. It conducts a focused interview, infers the project’s visual language, generates five distinct UI variants, gathers interactive feedback, and outputs a final plan. All temporary artifacts are deleted when the session ends or the user cancels.

How this skill works

The skill auto-detects package manager, framework, and styling system, then scans the codebase for design tokens and common component patterns to infer colors, spacing, typography, and radii. It runs an interview to assemble a structured Design Brief, saves that brief, and scaffolds a temporary lab that renders five meaningfully different variants with an interactive FeedbackOverlay. When the user confirms a final design, the lab cleans up all temporary files and produces an implementation plan; if the user aborts, the lab removes all artifacts immediately.

When to use it

  • Explore multiple directions before committing to a single UI approach
  • Redesign an existing component or page and compare meaningful alternatives
  • Create a new component or page with clear tradeoff demonstrations
  • Validate interaction models and accessibility states with stakeholders
  • Collect lightweight, in-context user feedback during design exploration

Best practices

  • Always run the interview to produce a precise Design Brief before generating variants
  • Infer visual tokens from the codebase (Tailwind config, CSS variables, theme files) rather than using preset styles
  • Make each of the five variants explore a distinct design axis (hierarchy, layout, density, interaction, expressive)
  • Include a reliable FeedbackOverlay in the route directory so users can annotate and submit feedback
  • Ensure cleanup triggers on both confirm and cancel to never leave .claude-design or __design_lab behind

Example use cases

  • Redesign a checkout card and compare dense vs spacious layouts with clear tradeoffs
  • Prototype alternative interaction models for an inline edit flow vs modal edit flow
  • Create five landing page hero variants to test hierarchy and CTAs with stakeholders
  • Assess accessibility and motion choices across variants before implementation
  • Generate a final implementation plan with file-level guidance after stakeholder selection

FAQ

All temporary artifacts are deleted immediately; the lab confirms deletion and no plan is generated.

How are visual styles chosen for variants?

The skill reads project tokens (Tailwind, CSS variables, or UI theme files) and samples existing components to apply the project’s visual language to every variant.

Can I restrict technical constraints or dependencies?

Yes—the interview captures constraints (no new dependencies, reuse components, WCAG requirements) and the variants respect those constraints.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational