ams-xstate-pro
暂无简介
- License 许可
- MIT
- 最近更新
- 2026-11-07
- Stars
- ★ 0
- Forks
- 1
03 SKILL.md 正文
XState Pro
A practical guide to XState (v5): modeling application logic as state machines and statecharts — explicit states, events, transitions, guards, and actions — plus the actor model for communicating machines. For anywhere boolean flags (isLoading, isError, isOpen...) multiply into impossible combinations.
Overview
Most UI bugs are impossible states: isLoading && isError, a form that's both submitting and editable, a player that's playing and paused. State machines eliminate them by construction: the machine is in exactly one state, and only declared transitions move it. XState adds hierarchy (nested states), parallelism (orthogonal regions), guards (conditional transitions), and actions (side effects) — the full statechart formalism — with TypeScript inference over states and events.
When to use
- Multi-step flows: wizards, checkout, onboarding, auth (idle → loading → success/error).
- Complex components: data fetching with retry, media players, editors with modes.
- Workflows and sagas: order processing, background jobs with states.
- Replacing sprawling
useReduceror flag-soupuseState. - Visualizing/debugging logic (machines can be diagrammed and model-tested).
Core concepts
- Machine.
setup({ types, actions, guards }).createMachine({ id, initial, context, states })— states,on: { EVENT: target }transitions,context(extended state: data, not control state). - Events.
{ type: 'SUBMIT' }(+ payload). Everything that happens is an event — user actions, timers, promise resolutions. Never mutate; send events. - Transitions.
on: { RETRY: 'loading' }, guarded:SUBMIT: { guard: 'isValid', target: 'submitting' }, with actions:actions: 'notifyUser'. - Actions.
assign(update context),sendTo/spawning actors, custom actions. Actions are declarative descriptions — testable without running effects. - Guards.
guard: ({ context, event }) => ...— conditions on transitions. Keep pure. - Hierarchical states.
editing: { initial: 'idle', states: { idle: {}, saving: {} } }— nested machines with inherited transitions. - Parallel states.
type: 'parallel'regions for orthogonal concerns (e.g.,ui: {...}+data: {...}simultaneously). - Actors.
createActor(machine)— running instances; actors communicate via messages.useActor/useMachinein React. - Invoked services.
invoke: { src: 'fetchUser', onDone: ..., onError: ... }— promises/observables as states, notuseEffectchains.
Practical workflow
1. Model before coding. Draw states on paper: what are the modes? What events move between them? What's impossible? The diagram is the design review.
2. Define the machine.
import { setup, assign } from 'xstate';
const formMachine = setup({
types: { context: {} as { error?: string }, events: {} as { type: 'SUBMIT' } | { type: 'RETRY' } },
guards: { isValid: ({ event }) => validate(event) },
}).createMachine({
id: 'form',
initial: 'editing',
states: {
editing: { on: { SUBMIT: { guard: 'isValid', target: 'submitting' } } },
submitting: {
invoke: { src: 'saveForm', onDone: 'success', onError: { target: 'editing', actions: assign({ error: ({ event }) => event.error.message }) } },
},
success: { type: 'final' },
},
});
3. Use in React.
import { useMachine } from '@xstate/react';
const [state, send] = useMachine(formMachine, { input: {...} });
state.matches('submitting'); // explicit, no flag soup
<button onClick={() => send({ type: 'SUBMIT' })} disabled={!state.can({ type: 'SUBMIT' })}>
4. Test the logic. Machines are pure data — assert transitions without rendering: actor.send({type:'SUBMIT'}); expect(actor.getSnapshot().value).toBe('submitting'). Model-based testing can even generate paths.
5. Visualize. Paste into the Stately visualizer during design reviews — stakeholders understand diagrams better than code.
Common pitfalls
- Context as control state.
context: { mode: 'loading' }duplicates what states express. States = control flow; context = data. If you're branching on context values, you probably need states. - Missing transitions. Unhandled events are silently ignored — usually correct, occasionally a bug. During dev, log unhandled events to catch typos.
- Actions with side effects inline.
assignand pure actions are testable; stuffing fetch calls into actions makes machines untestable. Side effects → invoked services or actors. - Over-modeling. A two-state toggle doesn't need a machine. Reach for XState when states × events gets beyond what you can hold in your head (~4+ states or tricky transitions).
- Guard impurity. Guards reading mutable externals make transitions unpredictable. Guards see
(context, event)— keep them pure. - Forgetting
input. v5 machines takeinputfor initialization; stuffing init data into context defaults couples machine to one use. - Parallel region explosion. Parallel states multiply combinations — use for genuinely orthogonal concerns, not as a grouping convenience.
- Not using
state.can(). Disabling buttons via ad-hoc flags reintroduces the flag soup.state.can(event)is the machine answering "is this allowed now?"
04 同仓库更多 Skills
web-scrape
Turn web pages into LLM-ready markdown via the Firecrawl API (renders JS, anti-bot handling). Use when browser_open/fetch fails on JS-heavy pages, or when you need clean markdown/structured extraction from a site. Requires FIRECRAWL_API_KEY.
design-taste-frontend
tasteskill v2, the anti-slop frontend skill. Art direction for websites, distilled from a study of 247 curated gallery sites (One Page Love, Awwwards Sites of the Day). Deals researched directions (idea, composition, type, palette, photographs, section blueprints) from a library, enforces a strict no-slop house style with a real-browser inspector, and runs a look-and-fix loop on screenshots. Use this whenever you build, design, redesign, restyle, polish or review a landing page, marketing site, portfolio, product page, one-pager, hero or any public-facing web page, in any stack (HTML/CSS, React, Next.js, Vue, Svelte, Astro, Tailwind) — also when the user says a page looks generic, templated, "AI-made", sloppy, or asks to make it look designed, premium, award-level or less like every other site. Not for dashboards or app screens.
test-driven-development
Use when implementing any feature or bugfix, before writing implementation code
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes
requesting-code-review
Use when completing tasks, implementing major features, or before merging to verify work meets requirements
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).