AGENT SKILL

shiny-for-r

Building, styling, testing, debugging, or observing a Shiny for R app - library(shiny), shinyApp(), runApp(). Index skill: read this, then open the linked reference for the task. Covers reactivity (reactive/observe/req/isolate/bindEvent); modules; async promises and ExtendedTask; bslib layouts, navigation, cards, value boxes, and theming (incl. thematic and brand.yml); dynamic UI; plots with click/brush interaction; tables, file upload/download; notifications, modals, progress, validation; bookmarking; custom JS components; testServer testing; debugging; OpenTelemetry; and when to reach for DT, plotly, leaflet, shinychat, or shinytest2. Use when writing or changing any Shiny for R app, especially a dashboard, or when tempted to hand-roll what the framework provides - custom HTML tables, fake tabs, polling loops, blocking reactive work, or print-debugging server state.

License 许可
MIT
所属仓库
MichalLauer/skills
最近更新
2026-11-06
Stars
★ 0
Forks
0

01 前言 Front Matter

name
shiny-for-r
license
MIT
metadata
{ "author": "Posit Software, PBC", "version": "1.0" }
description
Building, styling, testing, debugging, or observing a Shiny for R app - library(shiny), shinyApp(), runApp(). Index skill: read this, then open the linked reference for the task. Covers reactivity (reactive/observe/req/isolate/bindEvent); modules; async promises and ExtendedTask; bslib layouts, navigation, cards, value boxes, and theming (incl. thematic and brand.yml); dynamic UI; plots with click/brush interaction; tables, file upload/download; notifications, modals, progress, validation; bookmarking; custom JS components; testServer testing; debugging; OpenTelemetry; and when to reach for DT, plotly, leaflet, shinychat, or shinytest2. Use when writing or changing any Shiny for R app, especially a dashboard, or when tempted to hand-roll what the framework provides - custom HTML tables, fake tabs, polling loops, blocking reactive work, or print-debugging server state.

02 资源文件 21

03 SKILL.md 正文

Shiny for R

A Shiny app has a ui object (or function), a server(input, output, session) function, and shinyApp(ui, server) to run them together. The reactive graph is the engine: reading a reactive source (input$x, a reactive(), a reactiveVal()) inside a reactive context registers a dependency, so changing that source re-runs everything that read it — you never call outputs or schedule updates yourself.

This skill is an index. Find your task below and read the linked reference file before writing code for that area. Read the one or two references your task actually touches, not all of them.

Not every corner of Shiny has a reference here — plain inputs (selectInput(), actionButton(), ...), deployment, and app performance tuning are not covered. For those, work from your own knowledge of the package and its help pages.

Foundations

Topic Use when Reference
Reactivity A value should recompute or an output update as inputs change; choosing between reactive()/observe()/observeEvent(); req(), isolate(), bindEvent(), timers, polling references/reactivity.md
Modules A reusable, repeatable UI+server component; avoiding input/output id collisions across copies references/modules.md
Session lifecycle Per-session cleanup (onSessionEnded()), flush hooks, unhandled errors, per-session vs app-level scope, session$userData references/session-lifecycle.md

Async

Topic Use when Reference
Async Returning a promises::promise from a render/observer so R keeps serving other sessions during slow I/O references/async.md
Extended tasks Running slow work in a background process while the session that started it stays fully interactive; task buttons references/extended-tasks.md

Dashboard building

Topic Use when Reference
Dashboard design Turning a dataset or brief into a polished analytical dashboard; information hierarchy, shared filters, responsive layout, empty states, final visual/functional pass references/dashboard-design.md
Dashboard components Composing bslib card(), value_box(), accordion(), tooltip()/popover() instead of hand-rolled div() markup references/dashboard-components.md
Layouts Arranging a page with bslib page_sidebar(), layout_columns(), card() instead of nested fluidRow()/column() pyramids references/layouts.md

Navigation & dynamic UI

Topic Use when Reference
Navigation Grouping switchable content into a real tab container/navbar instead of faking tabs with buttons plus conditionalPanel() references/navigation.md
Dynamic UI UI that must change after render — update*Input(), conditionalPanel(), renderUI()/uiOutput(), or insertUI()/removeUI() references/dynamic-ui.md

Outputs

Topic Use when Reference
Plots Pairing renderPlot()/plotOutput() for base/lattice/ggplot2 graphics; click/hover/brush interaction via nearPoints()/brushedPoints() references/plots.md
Tables Choosing renderTable() (static), the deprecated renderDataTable(), or DT/reactable for sortable/searchable interactive tables references/tables.md
Files fileInput() uploads and downloadHandler()/downloadButton() generated-file downloads references/files.md

Feedback & state

Topic Use when Reference
Feedback Toasts/notifications, modal dialogs, progress bars, and input validation sent from the server, typically inside observeEvent() references/feedback.md
Bookmarking Saving/restoring app state via a shareable URL instead of a hand-built query string or custom persistence layer references/bookmarking.md

Theming, assets & extending

Topic Use when Reference
Theming & assets Setting bslib::bs_theme() Sass variables, thematic for plot theming, brand.yml, or adding custom CSS/JS/image assets references/theming-assets.md
Custom components Integrating a third-party JS widget or bespoke interaction when no built-in input/output covers it references/custom-components.md

Testing & observability

Topic Use when Reference
Testing Verifying server-side reactive logic with testServer() without starting a real Shiny process or browser references/testing.md
Debugging Dropping into a real R debugger (browser(), error breakpoints) instead of print()-debugging a running reactive graph references/debugging.md
OpenTelemetry Tracing sessions, flush cycles, and reactive/observer/output executions instead of log-scraping or ad hoc timing references/opentelemetry.md
Ecosystem Deciding whether a companion package (DT, plotly, leaflet, shinychat, shinytest2, etc.) already solves the problem before hand-rolling it references/ecosystem.md

04 同仓库更多 Skills