ui-ux-design
Skill1 starsChanged 3 months agoMITView on GitHub
Designs a user interface's flow, states, hierarchy, responsive behaviour, accessibility, content, and visual direction before it is built. Use when a new or revised interface still has open design decisions, or when the user says design this screen, asks what it should look like, or how the user gets through it. Skip backend-only work and cosmetic edits whose direction is already decided.
20 other projects use a version of this.
Adopted by
Most recently taken up by
- jDxve/avengers-kittheir own version, first seen Nov 8, 2026
- mblauberg/provenanttheir own version, first seen Nov 8, 2026
- SWEStash/swe-workflow-skillstheir own version, first seen Nov 8, 2026
- delorenj/skillstheir own version, first seen Nov 7, 2026
- velonx/agent-skillstheir own version, first seen Nov 1, 2026
- rogue-dev-studio/ai-agents-roguetheir own version, first seen Nov 1, 2026
Used with
Projects with it also use prd-generator and security-review.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
dhzrxSkillprd-generator
Generate comprehensive Product Requirements Documents with AI PM best practices for new features and products
Also in 3 projects that use this
SentrySkillsecurity-review
Finds exploitable application security vulnerabilities in code changes. Use for Warden security scans, appsec review, OWASP-style checks, authentication or authorization bugs, injection, XSS, SSRF, path traversal, secrets, unsafe crypto, webhook verification, open redirects, or sensitive data exposure.
Also in 3 projects that use this
Other versions20 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 20
- EmeaAppGbb/spec2cloud★ 100same name
- felvieira/claude-skills-fv★ 28same name
- arvindand/agent-skills★ 17same name
- delorenj/skills★ 17same name
- crgarcia12/Liliput★ 4same name
- hahuyhungdev/ai-coding-config★ 3same name
- alex-voloshin-dev/ai-skills★ 2same name
- AntJanus/skillbox★ 2same name
- rogue-dev-studio/ai-agents-rogue★ 2same name
- hicay/claude-code-skills★ 1same name
10 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository50 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- working-from-a-ticketSkill
- writing-specsSkill
- clarifying-intentSkill
- dispatching-parallel-agentsSkill
- handoffSkill
- mapping-the-codebaseSkill
- prototypingSkill
- runtime-evidenceSkill
- using-git-worktreesSkill
- using-sdlc-skillsSkill
- viewing-artifactsSkill
- writing-skillsSkill
- yagniSkill
- finishing-a-branchSkill
- release-readinessSkill
- architecture-decisionsSkill
- coding-standardsSkill
- data-modelSkill
- migration-strategySkill
- reviewing-plansSkill
- system-architectureSkill
- writing-plansSkill
- adding-observabilitySkill
- changing-a-live-schemaSkill
- executing-plansSkill
- rollout-controlsSkill
- subagent-driven-developmentSkill
- test-driven-developmentSkill
- complexity-auditSkill
- containing-an-incidentSkill
- debuggingSkill
- deprecating-an-apiSkill
- diagnosing-a-sessionSkill
- eliminating-correctionsSkill
- large-scale-changesSkill
- post-mortemSkill
- refactor-architectureSkill
- brainstormingSkill
- define-goalsSkill
- feasibility-checkSkill
- scopingSkill
- receiving-code-reviewSkill
- requesting-code-reviewSkill
- reviewing-a-prSkill
- security-auditsSkill
- verification-before-completionSkill
- verification-strategySkill
- visual-ui-verificationSkill
The file
We split this skill into the rules it gives an agent
10 rules. 0 are common, 3 are distinctive, and 7 are very specific to this project.
0
Common
Most of the 17 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
7
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 51
If the visual direction is open, read references/design-quality.md before defining layout, type, color, spacing, shape, imagery, and motion …
- Very specific: line 55
If a visual, spatial, or motion decision is open, read references/visual-decisions.md before building the comparison surface, then build it …
- Distinctive: line 71
If serving fails or is declined, the file path stands as what was delivered.
- Very specific: line 88
If preferences selected rendered variants, freeze the keyed Selected visual references collection exactly as visual-decisions.md defines it …
- Very specific: line 91
Write the whole section to .sdlc-skills/designs/{{YYYY-MM-DD}}-{{topic}}.md or the user-set path, preserving approved sections around it.
- Very specific: line 109
Ask this and end the turn.
- Distinctive: line 110
If a successor section replaces this one, its delta (the template's *Stable ID delta*) includes the selected visual references.
- Distinctive: line 112
If option 1 is chosen and every design section the work needs is approved,
- Very specific: line 126
Run without --root, bash scripts/start-server.sh exits 1, and the file-path fallback in Step 3.3 then hides that no surface was served.
- Very specific: line 128
Run without --entry, the printed URL opens the visuals directory.
“Common” is measured: kept by at least 6 in 10 of the 17 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
UI/UX Design
In an existing product, the current system is evidence: extend it unless the brief explicitly calls for change.
When to use
- A new or revised flow, screen, or interface leaves behavior, hierarchy, or visual direction open.
- Skip for backend-only or non-interactive work, and for an exact cosmetic edit whose direction is already fixed.
- Scale down: a small, settled interface may need only one flow, its states, its hierarchy, and its acceptance checks; alternatives are not mandatory.
Available scripts
scripts/start-server.shandscripts/stop-server.sh— start and stop the key-gated localhost preview of a comparison surface (a per-session key, an owner watchdog, and an idle timeout). They wrapscripts/serve.py; read visual-decisions.md before starting or stopping a preview; it owns the details.
Step 1: Read the real context
Open assets/ui-ux-section.md before the steps below fill it in.
- Establish the requirements and the audience.
- In an existing project, find its routes, screens, components, tokens, type and content patterns, preview tooling, responsive and accessibility conventions, and tests.
- Sort what you found into three buckets: deliberate constraints that bind
this work, inconsistencies this work may correct, and
unknown, a pattern nobody can show was decided; ask before binding or correcting it. - In an existing product with no evidence reachable, stop. Ask for the repository, a preview, screenshots, or the system documentation. Never put a generic system in its place.
Step 2: Frame, flow, hierarchy
- Fill the template's Frame, Flows, State matrix, and Hierarchy and content sections, following the note on each.
Step 3: Decide the direction by seeing it
- If the visual direction is open, read
references/design-quality.mdbefore defining layout, type, color, spacing, shape, imagery, and motion as one product-specific system. While the hierarchy is unresolved, add no decoration around it. - If a visual, spatial, or motion decision is open, read
references/visual-decisions.mdbefore building the comparison surface, then build it on the first medium in the reference's order that fits. A self-contained local page starts fromassets/comparison-template.htmlbefore its first variant is written. Give every version block and every variant a stable ID. Say where the surface is, by its medium: a project preview by the route or command the project already has, a private page by its location, screenshots or wireframes by their paths, and a local page by its file path with an offer to serve it. Render a local page and check it at 360px first; with no way to render it, say so. Start the preview only after the user accepts, then present its URL:
bash scripts/start-server.sh --root .sdlc-skills/designs/{{YYYY-MM-DD}}-{{topic}}/visuals --entry {{decision-slug}}.htmlIf serving fails or is declined, the file path stands as what was delivered. Then ask which variant and why, and end the turn; never choose it yourself.
- Skip the surface only with a recorded reason, which must be one of these:
every open question is conceptual (its answer does not depend on seeing
anything); the scale-down clause under When to use applies; or the
uncertainty is feasibility, in which case invoke
prototyping. - Fill the template's Conditions table for all eight families. For each family not answered, write a skip record with an accountable owner.
Step 4: Classify evidence, compile, present
- Tag every claim with its kind in the template's Evidence table. A stakeholder's preference selects a direction; it never proves that users can complete the flow. If a usability risk is open, record it as open with a named check (the template's Future evaluator: what would settle it) and an owner.
- If preferences selected rendered variants, freeze the keyed Selected
visual references collection exactly as
visual-decisions.mddefines it before compiling. Memory, a path, or a label is not a contract. - Write the whole section to
.sdlc-skills/designs/{{YYYY-MM-DD}}-{{topic}}.mdor the user-set path, preserving approved sections around it. Selected screens, a chosen variant, or an agreed flow are inputs to the section, not the section itself. Present:
{{Section}} {{path}} — version {{identity (per template)}} {{summary lines}} 1. Approve and hand off to planning 2. Request changes 3. Reject 4. Cancel Recommendation: {{option}} — {{one sentence}}.Ask this and end the turn;
clarifying-intentowns what closes it.- If a successor section replaces this one, its delta (the template's Stable ID delta) includes the selected visual references.
- If option 1 is chosen and every design section the work needs is approved,
REQUIRED SUB-SKILL: invoke
writing-plansagainst this version. If a needed section is not approved yet, invoke its owning skill instead.
Gotchas
- One open question can have a conceptual half and a visual half: whether an error should block submission, and separately whether that error should look alarming or subtle. The skip test in Step 3.4 asks whether every open question is conceptual, so settling the conceptual half in prose does not release the visual half from being rendered and compared.
- A pattern that is neither a deliberate constraint nor an inconsistency is
an undocumented accident nobody decided. Filing it as deliberate by default
treats an accident as a constraint; Step 1.3's
unknownbucket holds it. - Run without
--root,bash scripts/start-server.shexits 1, and the file-path fallback in Step 3.3 then hides that no surface was served. - Run without
--entry, the printed URL opens the visuals directory; that directory holds{{decision-slug}}.htmlpages and noindex.html, so the link the user gets returns 404. - A design request is not consent to a background listener: a preview started unasked opens a local port the user never agreed to.
Common mistakes
- Combining screens that were each selected on their own into a journey nobody reviewed.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

