design-patterns
/design-patterns
Replace generic interface choices with patterns justified by the user's task, content and product constraints. Use for a visual redesign, a UI critique or a request to make a screen less templated; excludes backend architecture and software design patterns.
- Author
- DroBov1Tya ↗
- Repository
- dusk_code ↗
- Licence
- MIT
- Last repository change
- 13.11.2026
Assessment and security
- Files in this package
- 1
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 13.11.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Replace generic interface choices with patterns justified by the user's task, content and product constraints. Use for a visual redesign, a UI critique or a request to make a screen less templated; excludes backend architecture and software design patterns.
Original author description
Replace generic interface choices with patterns justified by the user's task, content and product constraints. Use for a visual redesign, a UI critique or a request to make a screen less templated; excludes backend architecture and software design patterns.
How it works
- Establish the decision surface
- Keep a small defect ledger
- Match the pattern to the work
- Express hierarchy consistently
- Judge the result in use
SKILL.md preview
--- name: design-patterns description: Replace generic interface choices with patterns justified by the user's task, content and product constraints. Use for a visual redesign, a UI critique or a request to make a screen less templated; excludes backend architecture and software design patterns. source: "dusk_code" author: "Neil Ermak" license: MIT --- # Design Patterns Make the interface explain its purpose through its organization and behavior. An unusual font or a different accent cannot repair a screen whose structure hides the work people came to do. Choose patterns by the information and decisions they support, then give those patterns a coherent visual language. ## Establish the decision surface Identify the screen, intended audience, most frequent task, and the consequence of an incorrect action. Read the existing components and theme before proposing a replacement. Preserve the user's chosen style, content, workflows and public interfaces unless changing them belongs to the request. Use real content lengths and representative data. For a monitoring console, ask which signal deserves investigation; for a catalog, which attributes distinguish items; for a form, what information a person needs before committing. These answers determine prominence and grouping more reliably than a trend board. If the input is an image or a public URL without editable source, separate a critique or proposed implementation from changes to the original product. A reconstruction is a new artifact and cannot establish the original's behavior. ## Keep a small defect ledger For each material problem record: location, visible symptom, effect on a task, proposed correction, and how it can be checked. Prioritize lost information, ambiguous actions and inaccessible controls before decorative differences. Distinguish an observed defect from a preference or an untested hypothesis. Treat repeated card containers, competing accents, uniform type emphasis and ornamental labels as questions to investigate. They are not automatic defects. A familiar component is appropriate when it helps people recognize an action. If the current design works, report that finding and limit edits to demonstrated problems rather than manufacturing a redesign. ## Match the pattern to the work | Information relationship | Candidate arrangement | Check before choosing | |---|---|---| | Many records compared by the same attributes | Aligned rows with shared headings | Can people scan one attribute without opening each record? | | One decision requires supporting context | Main workspace beside an inspector | Does the selected item remain visible while details change? | | A long task has dependent inputs | Sections with a persistent completion summary | Can people revise earlier input without losing later work? | | Events must be interpreted over time | Time-based grouping with explicit units | Are gaps, ordering and time zones unambiguous? | | A visitor needs to understand an unfamiliar service | Concrete demonstration beside a concise explanation | Is the demonstration truthful and readable at the intended size? | These are candidates, not a mandatory template collection. Compare alternatives only where the tradeoff matters. Record the chosen organization, the content that earns emphasis, and the constraints it must survive. Proceed within existing authorization; ask only when a missing decision would materially change scope. ## Express hierarchy consistently Derive spacing from content groups and repeated alignment anchors. Use the project's tokens where possible; introduce a new token only for a repeated role. Distinguish body text, navigation, headings and quantitative values through their reading needs. Verify line wrapping with actual strings, including localization. Assign color a job: selection, status, grouping or brand recognition. A status must remain understandable without color alone. Choose imagery for explanatory value and available rights. Do not add a remote font, asset service, dependency or generated image merely to make the result look different. Write action labels around the resulting user operation. Keep the same operation named consistently across controls and feedback. Preserve useful context in empty, pending, failed and completed states. Animate only when the transition helps someone understand a change, and respect reduced-motion preferences. Apply the smallest coherent set of changes that resolves the ledger. Reuse existing components and interaction contracts. Avoid broad framework or component rewrites as a side effect of visual work. ## Judge the result in use When execution is available, inspect the rendered screen at representative wide and narrow sizes. Walk the primary task with keyboard and pointer, check focus visibility, and exercise relevant loading, empty and error states. Verify zoom, overflow and long content where they can affect the changed screen. For terminal interfaces, inspect the actual terminal at representative dimensions; a browser preview does not prove terminal input or layout behavior. Compare each ledger item with the resulting screen and its interaction. Run the project checks appropriate to the implementation. If a browser, terminal, service or test account is unavailable, state the unverified behavior specifically; source inspection and compilation do not establish visual or interaction quality. Deliver the changed artifacts, the reasoning behind the consequential choices, the observed checks, and unresolved ledger items. Include useful before/after captures when available without exposing private data. Stop when the requested outcome is evidenced; extra decoration is not an acceptance criterion.
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · DroBov1Tya/dusk_code | 95 | MIT |
| spencerbeggs/effected | 100 | MIT |
| spencerbeggs/effected | 100 | MIT |
| joeblackwaslike/agent-skills | 100 | MIT |
| krzysztofsurdy/code-virtuoso | 100 | MIT |
| nimadorostkar/Claude-Skills-collection | 100 | MIT |
| andreibesleaga/gabbe | 100 | CC-BY-4.0 |
| pantheon-org/tekhne | 100 | MIT |
| itzcull/software-skills | 100 | MIT |
Database snapshot: 13.11.2026 · package version aca98679fcd4 (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100ui-design
Unify product UI design, frontend implementation, and verification. Use when asked to design, redesign, build, or fix a web, desktop, or mobile interface; a page, dashboard, form, component, design system, responsive layout, accessibility, interface animation, or editable… - 100design-quality
Use for ANY work that changes how a UI looks: a new app/site/screen, a theme or redesign, a new logo or app icon, a new component, or even a small UI fix inside an existing product (player, reader, dashboard). Two modes: PRESERVE (editing an existing product — keep its design… - 100interface-refinement
Refine an existing interface with named commands (critique, polish, distill, typeset, colorize ...), a product and design context file, and a static detector of generic UI patterns. Use when UI looks generic, cluttered or unfinished. - 100ai-design-tells
Identify and fix generic AI design patterns in websites, landing pages, portfolios, and product interfaces. Use for requests about AI design tells, AI slop, a vibe-coded appearance, templated layouts, or preventing generic design during a build. Supports visual audits, focused… - 100screen-flow-patterns
Taxonomy of 25+ screen types, 15+ user flows, and 25+ UI element deep-dives with layout patterns, component hierarchies, state matrices, and best-in-class references. Use when deciding which screens a product needs and what connects them, before any visual design starts. - 90product-interface-design
Use when product feature design, multi-screen experience or high-fidelity interaction specification. Translate flows and constraints into polished implementable UI screens and interactions. Do not use when only writing backend code with no UI impact.