design-shiny-ui
About
This skill helps developers build modern, responsive Shiny app UIs using bslib for theming, layout_columns for grids, and components like value boxes and cards. It's ideal for creating new apps from scratch, modernizing legacy fluidPage apps, or applying brand theming and accessibility improvements. The skill also covers custom CSS/SCSS integration for complete design control.
Quick Install
Claude Code
Recommendednpx skills add pjt222/agent-almanac -a claude-code/plugin add https://github.com/pjt222/agent-almanacgit clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/design-shiny-uiCopy and paste this command in Claude Code to install this skill
Documentation
Design Shiny UI
Responsive, accessible Shiny UI. bslib theming, modern layout primitives, custom CSS.
When Use
- Build new Shiny app UI from scratch
- Modernize existing Shiny app: fluidPage → bslib
- Apply brand theming (colors, fonts) to Shiny app
- Make Shiny app responsive across screen sizes
- Improve accessibility
Inputs
- Required: App purpose + target audience
- Required: Layout type (sidebar, navbar, fillable, dashboard)
- Optional: Brand colors + fonts
- Optional: Use custom CSS/SCSS? (default: bslib only)
- Optional: Accessibility needs (WCAG level)
Steps
Step 1: Pick Page Layout
bslib has several page constructors:
# Sidebar layout — most common for data apps
ui <- page_sidebar(
title = "My App",
sidebar = sidebar("Controls here"),
"Main content here"
)
# Navbar layout — for multi-page apps
ui <- page_navbar(
title = "My App",
nav_panel("Tab 1", "Content 1"),
nav_panel("Tab 2", "Content 2"),
nav_spacer(),
nav_item(actionButton("help", "Help"))
)
# Fillable layout — content fills available space
ui <- page_fillable(
card(
full_screen = TRUE,
plotOutput("plot")
)
)
# Dashboard layout — grid of value boxes and cards
ui <- page_sidebar(
title = "Dashboard",
sidebar = sidebar(open = "closed", "Filters"),
layout_columns(
fill = FALSE,
value_box("Revenue", "$1.2M", theme = "primary"),
value_box("Users", "4,521", theme = "success"),
value_box("Uptime", "99.9%", theme = "info")
),
layout_columns(
card(plotOutput("chart1")),
card(plotOutput("chart2"))
)
)
Got: Page layout matches app's navigation + content needs.
If fail: Layout wrong? Check you use page_sidebar() / page_navbar() (bslib), not fluidPage() / navbarPage() (base shiny). bslib has better defaults + theming.
Step 2: Config bslib Theme
my_theme <- bslib::bs_theme(
version = 5, # Bootstrap 5
bootswatch = "flatly", # Optional preset theme
bg = "#ffffff", # Background color
fg = "#2c3e50", # Foreground (text) color
primary = "#2c3e50", # Primary brand color
secondary = "#95a5a6", # Secondary color
success = "#18bc9c",
info = "#3498db",
warning = "#f39c12",
danger = "#e74c3c",
base_font = bslib::font_google("Source Sans Pro"),
heading_font = bslib::font_google("Source Sans Pro", wght = 600),
code_font = bslib::font_google("Fira Code"),
"navbar-bg" = "#2c3e50"
)
ui <- page_sidebar(
theme = my_theme,
title = "Themed App",
# ...
)
Use interactive theme editor during dev:
bslib::bs_theme_preview(my_theme)
Got: App renders with consistent brand colors, fonts, Bootstrap 5 components.
If fail: Fonts don't load? Check internet (Google Fonts needs it) or switch to system fonts: font_collection("system-ui", "-apple-system", "Segoe UI"). Theme vars don't apply? Check theme is passed to page func.
Step 3: Build Layout with Cards + Columns
ui <- page_sidebar(
theme = my_theme,
title = "Analysis Dashboard",
sidebar = sidebar(
width = 300,
title = "Filters",
selectInput("dataset", "Dataset", choices = c("iris", "mtcars")),
sliderInput("sample", "Sample %", 10, 100, 100, step = 10),
hr(),
actionButton("refresh", "Refresh", class = "btn-primary w-100")
),
# KPI row — non-filling
layout_columns(
fill = FALSE,
col_widths = c(4, 4, 4),
value_box(
title = "Observations",
value = textOutput("n_obs"),
showcase = bsicons::bs_icon("table"),
theme = "primary"
),
value_box(
title = "Variables",
value = textOutput("n_vars"),
showcase = bsicons::bs_icon("columns-gap"),
theme = "info"
),
value_box(
title = "Missing",
value = textOutput("n_missing"),
showcase = bsicons::bs_icon("exclamation-triangle"),
theme = "warning"
)
),
# Main content row
layout_columns(
col_widths = c(8, 4),
card(
card_header("Distribution"),
full_screen = TRUE,
plotOutput("main_plot")
),
card(
card_header("Summary"),
tableOutput("summary_table")
)
)
)
Key layout primitives:
layout_columns()— responsive grid withcol_widthscard()— content container, optional header/footervalue_box()— KPI display, icon + themelayout_sidebar()— nested sidebar inside cardsnavset_card_tab()— tabbed cards
Got: Responsive grid adapts to screen size.
If fail: Columns stack odd on wide screens? col_widths sum must = 12 (Bootstrap grid). Cards overlap? Set fill = FALSE on non-filling rows.
Step 4: Add Dynamic UI
server <- function(input, output, session) {
output$dynamic_filters <- renderUI({
data <- current_data()
tagList(
selectInput("col", "Column", choices = names(data)),
if (is.numeric(data[[input$col]])) {
sliderInput("range", "Range",
min = min(data[[input$col]], na.rm = TRUE),
max = max(data[[input$col]], na.rm = TRUE),
value = range(data[[input$col]], na.rm = TRUE)
)
} else {
selectInput("values", "Values",
choices = unique(data[[input$col]]),
multiple = TRUE
)
}
)
})
# Conditional panels (no server round-trip)
# In UI:
# conditionalPanel(
# condition = "input.show_advanced == true",
# numericInput("alpha", "Alpha", 0.05)
# )
}
Got: UI updates dynamically from user selections + data.
If fail: Dynamic UI flickers? Use conditionalPanel() (CSS-based) not renderUI() where possible. Dynamic inputs lose values on re-render? Add session$sendInputMessage() to restore state.
Step 5: Custom CSS/SCSS (Optional)
For styles beyond bslib theme vars:
# Inline CSS
ui <- page_sidebar(
theme = my_theme,
tags$head(tags$style(HTML("
.sidebar { border-right: 2px solid var(--bs-primary); }
.card-header { font-weight: 600; }
.value-box .value { font-size: 2.5rem; }
"))),
# ...
)
# External CSS file (place in www/ directory)
ui <- page_sidebar(
theme = my_theme,
tags$head(tags$link(rel = "stylesheet", href = "custom.css")),
# ...
)
SCSS integration with bslib:
my_theme <- bslib::bs_theme(version = 5) |>
bslib::bs_add_rules(sass::sass_file("www/custom.scss"))
Got: Custom styles applied. bslib theming not broken.
If fail: Custom CSS conflicts with bslib? Use Bootstrap CSS vars (var(--bs-primary)) not hardcoded colors. Theme changes then propagate.
Step 6: Ensure Accessibility
# Add ARIA labels to inputs
selectInput("category", "Category",
choices = c("A", "B", "C")
) |> tagAppendAttributes(`aria-describedby` = "category-help")
# Add alt text to plots
output$plot <- renderPlot({
plot(data(), main = "Distribution of Values")
}, alt = "Histogram showing the distribution of selected values")
# Ensure sufficient color contrast in theme
my_theme <- bslib::bs_theme(
version = 5,
bg = "#ffffff", # White background
fg = "#212529" # Dark text — 15.4:1 contrast ratio
)
# Use semantic HTML
tags$main(
role = "main",
tags$h1("Dashboard"),
tags$section(
`aria-label` = "Key Performance Indicators",
layout_columns(
# value boxes...
)
)
)
Got: App meets WCAG 2.1 AA: color contrast, keyboard nav, screen reader compat.
If fail: Test with browser dev tools accessibility audit (Lighthouse). Check color contrast with WebAIM's checker. All interactive elements keyboard-focusable.
Checks
- Page layout renders right on desktop + mobile widths
- bslib theme applies consistently
- Value boxes display right themes + icons
- Cards resize in responsive grid
- Custom CSS uses Bootstrap vars, not hardcoded values
- All plots have alt text
- Color contrast ≥ WCAG AA (4.5:1 for text)
- Interactive elements keyboard accessible
Pitfalls
- Mixing old + new Shiny UI: Don't mix
fluidPage()with bslib components. Usepage_sidebar(),page_navbar(), orpage_fillable()only. - Hardcoded colors in CSS: Use
var(--bs-primary)not#2c3e50. Hardcoded colors break when theme changes. - Missing
fill = FALSEon non-filling rows: Value box + summary rows usually shouldn't stretch. Setfill = FALSE. - Google Fonts in offline env: App deploys to air-gapped network? Use system fonts or self-hosted font files, not
font_google(). - Ignoring mobile: Test with browser responsive mode.
layout_columnsauto-stacks on narrow screens, custom CSS may not.
See Also
scaffold-shiny-app— initial app setup incl theme configbuild-shiny-module— modular UI componentsoptimize-shiny-performance— perf-conscious renderingreview-web-design— visual design review: layout, typography, colourreview-ux-ui— usability + accessibility review
GitHub Repository
Related Skills
content-collections
MetaThis skill provides a production-tested setup for Content Collections, a TypeScript-first tool that transforms Markdown/MDX files into type-safe data collections with Zod validation. Use it when building blogs, documentation sites, or content-heavy Vite + React applications to ensure type safety and automatic content validation. It covers everything from Vite plugin configuration and MDX compilation to deployment optimization and schema validation.
polymarket
MetaThis skill enables developers to build applications with the Polymarket prediction markets platform, including API integration for trading and market data. It also provides real-time data streaming via WebSocket to monitor live trades and market activity. Use it for implementing trading strategies or creating tools that process live market updates.
creating-opencode-plugins
MetaThis skill helps developers create OpenCode plugins that hook into 25+ event types like commands, files, and LSP operations. It provides the plugin structure, event API specifications, and implementation patterns for JavaScript/TypeScript modules. Use it when you need to intercept, monitor, or extend the OpenCode AI assistant's lifecycle with custom event-driven logic.
sglang
MetaSGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.
