CCM
/Skills
SkillsMCPMarketplacesDigestToolsAdvertise

This week in Claude

Every Monday: Claude Code, Agent SDK, MCP, and the Anthropic platform moves worth your time.

Skills by Category
Frontend DevelopmentBackend & APIsTesting & QASecurityDevOps & CI/CDGit & Pull RequestsDocumentationCode Review & QualityAI & Agent BuildingSkill Development
MCP Servers by Category
Sales & MarketingWeb & Browser AutomationDatabasesAI & LLM ToolsCloud & InfrastructureCommunication & MessagingDeveloper ToolsDesign & CreativeDocuments & KnowledgeSearch & Web Crawling
Marketplaces by Category
AI Agents & OrchestrationLLM IntegrationDevelopment ToolsFrontend & UIBackend & APIsDatabasesTesting & Code QualityDevOps & CloudSecurity & ComplianceGit & Version Control

Claude Code Marketplaces

Discover Claude Code plugins, extensions, and tools. Automatically updated directory of Anthropic Claude AI marketplaces with development tools, productivity plugins, and integrations.

Resources

  • Browse Skills
  • Browse MCP Servers
  • Browse Marketplaces
  • Skill index
  • MCP index
  • Marketplace index
  • Plugins Reference

Community

  • About
  • Tools
  • Feedback
  • Privacy Policy
  • Advertise

Built for the Claude Code community with Claude Code by mertbuilds.com

Independent project, not affiliated with Anthropic
tristanmanchester avatar

Designing Beautiful Websites

tristanmanchester/agent-skills
2.4k installs3 stars
Summary

This is a full UX/UI design workflow that takes you from strategy through to implementation specs. It enforces a 7-step process: gather inputs, define strategy, scope features, build information architecture, wireframe the skeleton, apply visual polish, then validate and hand off. The non-negotiables are good: reduce thinking, use conventions, clear hierarchy, unambiguous grouping. It includes a default design system with spacing scales, typography, and color rules that you can use immediately. Best for when someone asks you to design or redesign a website and you want structured deliverables that engineers can actually build from, not just pretty mockups.

Install to Claude Code

npx -y skills add tristanmanchester/agent-skills --skill designing-beautiful-websites --agent claude-code

Installs into .claude/skills of the current project.

CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
inference shell
inference shell
create and run specialised agents in minutes
build now →
MCP-ready Email SendingMCP-ready Email Sending
MCP-ready Email Sending
Plug Mailtrap into your AI workflow and let it handle the email.
Connect Mailtrap MCP →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Capacitor - Shared memory for your team’s coding agents.
Capacitor - Shared memory for your team’s coding agents.
Make coding agent sessions - Searchable, Shareable, Vendor-neutral & Scored.
Try For Free →
CodeScene MCP ServerCodeScene MCP Server
CodeScene MCP Server
Your agent targets a perfect 10 Code Health score. Deterministic. Every commit.
Try For Free →
Give your AI the whole web as clean markdownGive your AI the whole web as clean markdown
Give your AI the whole web as clean markdown
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
inference shell
inference shell
create and run specialised agents in minutes
build now →
MCP-ready Email SendingMCP-ready Email Sending
MCP-ready Email Sending
Plug Mailtrap into your AI workflow and let it handle the email.
Connect Mailtrap MCP →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Capacitor - Shared memory for your team’s coding agents.
Capacitor - Shared memory for your team’s coding agents.
Make coding agent sessions - Searchable, Shareable, Vendor-neutral & Scored.
Try For Free →
CodeScene MCP ServerCodeScene MCP Server
CodeScene MCP Server
Your agent targets a perfect 10 Code Health score. Deterministic. Every commit.
Try For Free →
Give your AI the whole web as clean markdownGive your AI the whole web as clean markdown
Give your AI the whole web as clean markdown
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
Files
SKILL.mdView on GitHub

Designing Beautiful Websites

Core philosophy: Make the next action obvious. Build from user goals upward. Systemise visuals. Validate early.

Why this exists

Websites fail when they look “nice” but:

  • don’t match user goals,
  • hide key actions,
  • require too much thinking,
  • or are visually inconsistent.

This skill turns vague requests like “make it look better” into a repeatable workflow that produces:

  • clear structure,
  • usable interactions,
  • and a cohesive visual system.

What “done” looks like

Deliverables should be usable by builders (engineers, no-code builders, future agents):

  • Design brief: users, goals, constraints, success metrics.
  • IA + flows: sitemap (or nav model), and 1–3 key user journeys.
  • Layout + wireframes: responsive page structure, component inventory.
  • Visual system: tokens (type, spacing, colour, radius, shadow), and usage rules.
  • Component specs: states, behaviour, empty/loading/error.
  • QA notes: accessibility, responsiveness, edge cases.

If time is limited, prioritise: clarity + hierarchy + consistency + accessibility.

Quick start workflow

Copy this checklist into the working notes and tick it off:

  • 0. Inputs: goal, audience, content, constraints, brand signals.
  • 1. Strategy: user goals + business goals + success metrics.
  • 2. Scope: pages/features/content; prioritise “key paths”.
  • 3. Structure: IA + navigation model + flows.
  • 4. Skeleton: wireframes + component inventory + responsive layout.
  • 5. Surface: visual system + page comps + states.
  • 6. Validate: usability pass + accessibility pass + consistency pass.
  • 7. Hand-off: tokens + component specs + implementation notes.

Default rule: do not jump to surface polish until structure and skeleton are believable.

Non‑negotiables

1) Reduce thinking

Design so users rarely wonder:

  • “Where am I?”
  • “What do I do next?”
  • “Is that clickable?”
  • “Why did they call it that?”

Prefer obvious over clever.

2) Use conventions aggressively

Use familiar patterns unless there is a measured reason to deviate. Unusual UI is a tax on every user interaction.

3) Clear visual hierarchy

Every screen must answer (at a glance):

  • what this page is,
  • what the primary action is,
  • where the navigation is,
  • what is secondary.

4) Grouping must be unambiguous

If spacing is doing grouping work:

  • there must be more space around groups than within groups.

5) Feedback and forgiveness

Users should:

  • see results of actions quickly,
  • understand system status,
  • and recover via undo/back/cancel where possible.

Prefer preventing errors over scolding users.

6) Accessibility is part of “beautiful”

Good aesthetics survive:

  • keyboard-only use,
  • low vision,
  • colour‑blindness,
  • small screens,
  • slow networks.

Default outputs format

When responding, produce:

  1. Design brief (bullets)
  2. IA + key flows (bullets + simple diagrams if useful)
  3. Component inventory (table or list)
  4. Design tokens (CSS variables or JSON)
  5. Page-level guidance (for each page/section)
  6. States & edge cases
  7. Implementation notes (HTML structure, CSS approach, ARIA, etc.)

If the user asked for a critique/audit, output:

  • issues (grouped by severity),
  • fixes,
  • and a “next iteration plan”.

The workflow in practice

Step 0 — Gather inputs (fast)

Ask only what’s needed; otherwise assume and state assumptions.

Minimum questions:

  • Primary user: who is this for?
  • Primary goal: what must they do?
  • Business goal: what does success look like?
  • Content: what is real copy/data?
  • Brand signals: existing colours/logo/type/voice?
  • Constraints: tech stack, deadline, accessibility level.

If inputs are missing, create a working brief with explicit assumptions.

Step 1 — Strategy (align intent)

Produce:

  • primary + secondary user goals,
  • business objectives,
  • success metrics,
  • constraints/risk.

Step 2 — Scope (decide what exists)

Define:

  • pages/screens,
  • features,
  • content requirements,
  • and what is out of scope.

Pick 1–3 key paths (the journeys that matter most). Optimise these first.

Step 3 — Structure (make it findable)

Create:

  • sitemap / nav model (global + local nav),
  • page purpose statements,
  • user flows for key paths.

Rule: navigation labels should be self‑evident; avoid internal jargon.

Step 4 — Skeleton (arrange the UI)

Create:

  • wireframes per page,
  • component inventory,
  • layout constraints (container widths, grids, spacing rhythm),
  • and priority order per breakpoint.

Rule: start with the feature/content, not the “app shell”.

Step 5 — Surface (make it beautiful)

Build a consistent system:

  • spacing + sizing scale,
  • typography scale,
  • colour palette + shades,
  • radius + border rules,
  • elevation/shadow scale,
  • icon + illustration style,
  • motion rules (optional).

Apply to page comps.

Step 6 — Validate (fast loops)

Run these checks:

  • Glance test (5–10 seconds): can someone tell what this is and what to do?
  • Key‑path walkthrough: can a first‑time user complete the main task?
  • Consistency pass: are tokens respected? is hierarchy consistent?
  • Accessibility pass: contrast, focus states, semantics, error messaging.

Step 7 — Hand-off (make it buildable)

Provide:

  • tokens,
  • component specs (states + spacing + behaviour),
  • responsive rules,
  • and edge cases.

Default starter system

Use this system unless the project already has one.

Spacing & sizing scale (px)

Use a non-linear scale so choices are easy: 0, 4, 8, 12, 16, 24, 32, 48, 64, 96, 128

Rules:

  • pick from the scale; avoid “one-off” numbers.
  • for grouping: inside-group spacing < between-group spacing.

Typography scale (px)

Keep it tight: 6–8 sizes is enough.

Suggested:

  • 12 caption
  • 14 small/body-secondary
  • 16 body
  • 20 subheading
  • 24 h3
  • 30 h2
  • 40 h1/hero

Rules:

  • default body line-height ~ 1.5–1.7.
  • limit line length for reading (~45–80 characters).
  • use weight/colour/spacing before adding new sizes.

Colour system

  • define neutrals (backgrounds + text), one primary, and semantic accents.
  • define shades up front (e.g., 100–900), don’t generate ad-hoc lightens/darkens.
  • “Grey” can be warm or cool; keep a consistent temperature.

Contrast rules:

  • normal text target: ≥ 4.5:1
  • large text target: ≥ 3:1

Elevation / shadow system

Use 3–5 shadow levels that map to meaning:

  • 1: buttons/cards (subtle)
  • 2: popovers/menus
  • 3: sticky headers
  • 4: modals
  • 5: high priority overlays

Borders

Prefer:

  • spacing,
  • subtle shadows,
  • or small background changes over heavy borders.

Empty states

Empty states are a first impression. They must:

  • explain what’s empty,
  • why it matters,
  • and what to do next.

Progressive disclosure

Use these reference files when deeper detail is needed:

  • Workflow & deliverables → references/WORKFLOW.md
  • Page patterns → references/PAGE-PATTERNS.md
  • Audit / critique → references/DESIGN-AUDIT.md
  • Usability & navigation → references/USABILITY.md
  • Visual design systems → references/VISUAL-DESIGN.md
  • Interaction & forms → references/INTERACTION-DESIGN.md
  • Information architecture → references/INFORMATION-ARCHITECTURE.md
  • Content & microcopy → references/CONTENT-COPY.md
  • Responsive rules → references/RESPONSIVE.md
  • Accessibility → references/ACCESSIBILITY.md
  • Checklists & templates → references/CHECKLISTS.md

Quick search

If running locally:

grep -i "empty state\|hierarchy\|spacing" -n references/*.md

THE EXACT PROMPT — UX/UI plan

You are designing a website UI/UX.

1) Write a crisp design brief (users, goals, constraints, success metrics).
2) Define information architecture + navigation model.
3) Identify 1–3 key user paths; write step-by-step flows.
4) Produce a component inventory for the key pages.
5) Propose a design token system (spacing, type, colour, radius, shadow) with rules.
6) Describe page layouts (mobile-first) and key interactions.
7) List empty/loading/error states and edge cases.
8) Run a usability + accessibility + consistency pass; revise.

Output must be specific and implementable.
Avoid vague advice.

THE EXACT PROMPT — Visual polish pass

Review this UI for visual quality.

- Fix hierarchy (what is primary vs secondary vs tertiary?)
- Fix spacing (grouping clarity; rhythm; alignment)
- Fix typography (scale, weights, line height, line length)
- Fix colour (contrast, palette consistency, accent usage)
- Fix depth (shadows/borders; focus on meaning)
- Improve empty states and microcopy

Return:
1) a list of concrete changes
2) updated tokens (if needed)
3) before/after descriptions of the most important screens.

THE EXACT PROMPT — Usability “glance test”

Pretend you have 10 seconds to look at this page.

Answer:
- What is this page?
- Who is it for?
- What are the top 3 things I can do here?
- What is the primary action?
- Where is the navigation?

Then list everything that created a question mark, and propose fixes.

THE EXACT PROMPT — Component spec (single component)

Write a build-ready spec for this component.

Include:
- Purpose + when to use
- Anatomy (parts)
- Variants
- States: default/hover/focus/active/disabled/loading/error/success/empty
- Behaviour rules (keyboard + mouse + touch)
- Spacing + typography + colour tokens used
- Accessibility notes (ARIA if needed)
- Edge cases (long text, missing data, localisation)

Keep it concise but unambiguous.

THE EXACT PROMPT — Accessibility pass

Review this design for accessibility.

Check:
- text contrast (normal and large text)
- keyboard navigation and focus visibility
- semantic element choices (button vs link vs div)
- form labelling and error announcement
- motion and reduced-motion behaviour

Return:
1) issues grouped by severity
2) concrete fixes (design + implementation)
3) any token changes needed (colours, focus styles)

THE EXACT PROMPT — Responsive pass

Define responsive behaviour for this page/component.

For each breakpoint (small phone, large phone, tablet, desktop):
- layout (stack/columns)
- what becomes primary vs secondary
- how text wraps/truncates
- how tables, toolbars, and secondary actions adapt

Then list edge cases (long text, empty, error) and how they render.
Featured
CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
inference shell
inference shell
create and run specialised agents in minutes
build now →
MCP-ready Email SendingMCP-ready Email Sending
MCP-ready Email Sending
Plug Mailtrap into your AI workflow and let it handle the email.
Connect Mailtrap MCP →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Capacitor - Shared memory for your team’s coding agents.
Capacitor - Shared memory for your team’s coding agents.
Make coding agent sessions - Searchable, Shareable, Vendor-neutral & Scored.
Try For Free →
CodeScene MCP ServerCodeScene MCP Server
CodeScene MCP Server
Your agent targets a perfect 10 Code Health score. Deterministic. Every commit.
Try For Free →
Give your AI the whole web as clean markdownGive your AI the whole web as clean markdown
Give your AI the whole web as clean markdown
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
Categories
Frontend DevelopmentDocumentationCode Review & QualityAI & Agent BuildingDesign & UI/UX
First SeenApr 16, 2026
View on GitHub

More from tristanmanchester/agent-skills

All 4 skills →
  • Reddit Readonly421
  • Styling Nativewind V4 Expo376
  • Animating React Native Expo369

Recommended

More Frontend Development →
nomadamas avatar
daishin-report-search

nomadamas/k-skill

대신증권 리포트 GitHub Pages 미러에서 최신 HTML 리포트 목록과 원문/설명 페이지를 조회한다.
2.4k
7.1k
deckardger avatar
tanstack-integration-best-practices

deckardger/tanstack-agent-skills

Best practices for integrating TanStack Query with TanStack Router and TanStack Start. Patterns for full-stack data flow, SSR, and caching coordination.
2.4k
205
microsoft avatar
update-screenshots

microsoft/vscode

Download screenshot baselines from the latest CI run and commit them. Use when asked to update, accept, or refresh component screenshot baselines from CI, or after the screenshot-test GitHub Action reports differences. This skill should be run as a subagent.
2.4k
188.5k
affaan-m avatar
coding-standards

affaan-m/ecc

Baseline cross-project coding conventions for naming, readability, immutability, and code-quality review. Use detailed frontend or backend skills for framework-specific patterns.
2.3k
240.8k
mantinedev avatar
mantine-combobox

mantinedev/skills

Build custom dropdown/select/autocomplete/multiselect components using Mantine's Combobox primitives. Use this skill when: (1) creating a new custom select-like component with Combobox primitives, (2) building a searchable dropdown, (3) implementing a multi-select or tags input variant, (4) customizing option rendering, (5) adding custom filtering logic, or (6) any task involving useCombobox, Combobox.Target, Combobox.Option, or Combobox.Dropdown.
2.3k
58
openai avatar
chatgpt-apps

openai/skills

Build, scaffold, refactor, and troubleshoot ChatGPT Apps SDK applications that combine an MCP server and widget UI. Use when Codex needs to design tools, register UI resources, wire the MCP Apps bridge or ChatGPT compatibility APIs, apply Apps SDK metadata or CSP or domain settings, or produce a docs-aligned project scaffold. Prefer a docs-first workflow by invoking the openai-docs skill or OpenAI developer docs MCP tools before generating code.
2.3k
24.7k