Design direction
Skill MaybeMonad/skills/skills/design-workflow/design-direction
npx -y skills add MaybeMonad/skills --skill design-directionAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Aesthetic direction workflow for greenfield UI, brandless screens, new visual systems, redesign concepts, or designs that need a concrete typography, color, density, component, imagery, and motion thesis before hi-fi work.
SKILL.md
3.5 KB, as published. Nobody here has run it
Design Direction
Use this skill when there is no design system to inherit from, or when the user explicitly wants a new look. Direction turns vague taste into a concrete visual contract before hi-fi work spreads across many screens.
Workflow
- Confirm the blank slate:
- Check for brand guides, tokens, codebase components, screenshots, Figma files, existing product UI, and reference products.
- Completion criterion: either an existing source is ruled out, or the work
switches to
design-source-map.
- Name the intent:
- Audience, product category, emotional tone, trust level, content density, novelty appetite, and explicit off-limits tropes.
- If the user is unsure, propose 2-3 distinct directions and ask them to pick or combine.
- Completion criterion: the direction can be summarized as a specific thesis, not "modern and clean."
- Commit the visual contract:
- Typography: families, weights, scale, line heights, and fallback plan.
- Color: primary, accent if needed, neutrals, semantic colors, surfaces, and contrast expectations.
- Density: spacing scale, layout rhythm, container widths, and information density.
- Shape and elevation: radii, borders, shadows, dividers, and focus rings.
- Components: button, input, card, nav, tab, menu, modal, table, and empty state direction where relevant.
- Imagery and icons: real assets, icon library, placeholders, or asset gaps.
- Motion: timing, easing, state transitions, entry/exit motion, and reduced motion behavior.
- Completion criterion: every axis has a concrete default or an explicit open decision.
- Prove it small:
- Build or specify one representative surface: hero, toolbar, card set, form, dashboard slice, mobile screen, or slide.
- Completion criterion: the proof surface exercises type, color, spacing, component style, and at least one interaction state.
- Verify the direction:
- Check contrast, text fit, responsive constraints, palette balance, hierarchy, and whether the surface reads as the intended thesis.
- Completion criterion: unresolved design judgments are named before the direction is reused.
Output Shape
- Direction thesis: audience, tone, and design promise.
- Token contract: type, color, spacing, shape, elevation, motion.
- Component contract: core UI defaults and state expectations.
- Proof surface: file, screenshot, or described slice.
- Open decisions: what the user should sign off before scaling the direction.
Rules
- Do not use a single-hue palette as the whole design unless the brand demands it; introduce restrained neutral and semantic structure.
- Do not default to decorative gradients, emoji, weak custom SVG scenes, or generic SaaS card patterns as the visual idea.
- Do not make a landing page when the user asked for an app, dashboard, tool, or game. Build the primary experience first.
- Add new values to the direction before using them in an artifact; avoid one-off inline styling.
- When the direction is for production code, map it to the project's token format instead of leaving it as prose.
Credit
Inspired by Trystan-SA/claude-design-system-prompt (MIT), adapted into a smaller direction-setting workflow.