diff --git a/README.md b/README.md
index 62d4f7d..c677389 100644
--- a/README.md
+++ b/README.md
@@ -1,311 +1,533 @@
 # boardgame-core
 
-A state management library for board games using [Preact Signals](https://preactjs.com/guide/v10/signals/).
+基于 [Preact Signals](https://preactjs.com/guide/v10/signals/) 的桌游状态管理库。
 
-Build turn-based board games with reactive state, entity collections, spatial regions, and a command-driven game loop.
+使用响应式状态、实体集合、空间区域和命令驱动的游戏循环来构建回合制桌游。
 
-## Features
+## 特性
 
-- **Reactive State Management**: Fine-grained reactivity powered by [@preact/signals-core](https://preactjs.com/guide/v10/signals/)
-- **Type Safe**: Full TypeScript support with strict mode and generic context extension
-- **Region System**: Spatial management with multi-axis positioning, alignment, and shuffling
-- **Command System**: CLI-style command parsing with schema validation, type coercion, and prompt support
-- **Game Lifecycle Management**: `GameHost` class provides clean setup/reset/dispose lifecycle for game sessions
-- **Deterministic RNG**: Seeded pseudo-random number generator (Mulberry32) for reproducible game states
+- **响应式状态管理**：基于 [@preact/signals-core](https://preactjs.com/guide/v10/signals/) 的细粒度响应
+- **类型安全**：完整的 TypeScript 支持，启用严格模式和泛型上下文扩展
+- **区域系统**：支持多轴定位、对齐和洗牌的空间管理
+- **命令系统**：CLI 风格的命令解析，带 schema 校验、类型转换和玩家输入提示
+- **游戏生命周期管理**：`GameHost` 类提供清晰的游戏设置/重置/销毁生命周期
+- **确定性 RNG**：Mulberry32 种子伪随机数生成器，用于可复现的游戏状态
 
-## Installation
+## 安装
 
 ```bash
 npm install boardgame-core
 ```
 
-## Writing a Game
+---
 
-### Defining a Game
+## 使用 GameHost
 
-Each game defines a command registry and exports a `createInitialState` function:
+`GameHost` 是游戏运行的核心容器，负责管理游戏状态、命令执行和玩家交互的生命周期。
+
+### 创建 GameHost
+
+通过 `createGameHost` 传入一个 GameModule 来创建：
 
 ```ts
-import {
-    createGameCommandRegistry,
-    Region,
-    createRegion,
-} from 'boardgame-core';
+import { createGameHost } from 'boardgame-core';
+import * as tictactoe from 'boardgame-core/samples/tic-tac-toe';
 
-// 1. Define your game-specific state
-type MyGameState = {
-    board: Region;
-    parts: Record<string, { id: string; regionId: string; position: number[] }>;
-    score: { white: number; black: number };
-    currentPlayer: 'white' | 'black';
-    winner: 'white' | 'black' | 'draw' | null;
-};
+const host = createGameHost(tictactoe);
+```
 
-// 2. Create initial state factory
-export function createInitialState(): MyGameState {
+### 响应式状态
+
+GameHost 暴露的所有属性都是响应式 Signal，可以直接用于 UI 渲染或 `effect()`：
+
+```ts
+import { effect } from '@preact/signals-core';
+
+// 游戏状态
+effect(() => {
+    console.log(host.state.value.currentPlayer);
+    console.log(host.state.value.winner);
+});
+
+// 生命周期状态: 'created' | 'running' | 'disposed'
+effect(() => {
+    console.log('Status:', host.status.value);
+});
+
+// 当前等待的玩家输入 schema
+effect(() => {
+    const schema = host.activePromptSchema.value;
+    if (schema) {
+        console.log('Waiting for:', schema.name, schema.params);
+    }
+});
+
+// 当前等待的玩家
+effect(() => {
+    console.log('Current player prompt:', host.activePromptPlayer.value);
+});
+```
+
+### 启动游戏
+
+调用 `setup()` 并传入初始化命令名来启动游戏：
+
+```ts
+await host.setup('setup');
+```
+
+这会重置游戏状态、取消当前活动提示、运行指定的 setup 命令，并将状态设为 `'running'`。
+
+### 处理玩家输入
+
+当命令通过 `this.prompt()` 等待玩家输入时，使用 `onInput()` 提交输入：
+
+```ts
+// 提交玩家操作，返回错误信息或 null
+const error = host.onInput('play X 1 2');
+
+if (error) {
+    console.log('输入无效:', error);
+    // 玩家可以重新输入
+} else {
+    // 输入已被接受，命令继续执行
+}
+```
+
+### 监听事件
+
+```ts
+// 监听游戏设置完成
+host.on('setup', () => {
+    console.log('Game initialized');
+});
+
+// 监听游戏销毁
+host.on('dispose', () => {
+    console.log('Game disposed');
+});
+
+// on() 返回取消订阅函数
+const unsubscribe = host.on('setup', handler);
+unsubscribe(); // 取消监听
+```
+
+### 重新开始游戏
+
+```ts
+// 取消当前命令，重置状态，重新运行 setup 命令
+await host.setup('setup');
+```
+
+### 销毁游戏
+
+```ts
+host.dispose();
+```
+
+销毁后会取消所有活动命令、清理事件监听器，并将状态设为 `'disposed'`。销毁后无法再次使用。
+
+### 完整示例
+
+```ts
+import { effect } from '@preact/signals-core';
+import { createGameHost } from 'boardgame-core';
+import * as tictactoe from 'boardgame-core/samples/tic-tac-toe';
+
+const host = createGameHost(tictactoe);
+
+// 监听状态变化
+effect(() => {
+    const state = host.state.value;
+    console.log(`${state.currentPlayer}'s turn (turn ${state.turn + 1})`);
+    if (state.winner) {
+        console.log('Winner:', state.winner);
+    }
+});
+
+// 启动游戏
+await host.setup('setup');
+
+// 游戏循环：等待提示 → 提交输入
+while (host.status.value === 'running' && host.activePromptSchema.value) {
+    const schema = host.activePromptSchema.value!;
+    console.log('Waiting for input:', schema.name);
+
+    // 这里可以从 UI/网络等获取输入
+    const input = await getPlayerInput();
+    const error = host.onInput(input);
+
+    if (error) {
+        console.log('Invalid:', error);
+    }
+}
+
+// 游戏结束后可以重新开始
+// await host.setup('setup');
+
+// 或彻底销毁
+// host.dispose();
+```
+
+---
+
+## 编写 GameModule
+
+GameModule 是定义游戏逻辑的模块，包含状态定义和命令注册。
+
+### GameModule 结构
+
+一个 GameModule 必须导出两个东西：
+
+```ts
+import { createGameCommandRegistry, createRegion } from 'boardgame-core';
+
+// 1. 定义游戏状态
+export function createInitialState() {
     return {
         board: createRegion('board', [
-            { name: 'x', min: 0, max: 5 },
-            { name: 'y', min: 0, max: 5 },
+            { name: 'x', min: 0, max: 2 },
+            { name: 'y', min: 0, max: 2 },
         ]),
-        parts: {},
-        score: { white: 0, black: 0 },
-        currentPlayer: 'white',
-        winner: null,
+        parts: {} as Record<string, Part>,
+        currentPlayer: 'X' as PlayerType,
+        winner: null as WinnerType,
+        turn: 0,
     };
 }
 
-// 3. Create registry and register commands
-const registration = createGameCommandRegistry<MyGameState>();
+// 2. 创建命令注册表并注册命令
+const registration = createGameCommandRegistry<ReturnType<typeof createInitialState>>();
 export const registry = registration.registry;
 
-registration.add('place <row:number> <col:number>', async function(cmd) {
-    const [row, col] = cmd.params as [number, number];
-    const player = this.context.value.currentPlayer;
+// 注册命令
+registration.add('setup', async function () {
+    // ... 命令逻辑
+});
 
-    // Mutate state via produce()
+registration.add('play <player> <row:number> <col:number>', async function (cmd) {
+    // ... 命令逻辑
+});
+```
+
+也可以使用 `createGameModule` 辅助函数：
+
+```ts
+import { createGameModule, createGameCommandRegistry, createRegion } from 'boardgame-core';
+
+export const gameModule = createGameModule({
+    registry: registration.registry,
+    createInitialState,
+});
+```
+
+### 定义游戏状态
+
+游戏状态是一个普通对象，通过 `createInitialState()` 工厂函数创建。建议使用 `ReturnType` 推导类型：
+
+```ts
+export function createInitialState() {
+    return {
+        board: createRegion('board', [
+            { name: 'x', min: 0, max: 2 },
+            { name: 'y', min: 0, max: 2 },
+        ]),
+        parts: {} as Record<string, TicTacToePart>,
+        currentPlayer: 'X' as PlayerType,
+        winner: null as WinnerType,
+        turn: 0,
+    };
+}
+export type GameState = ReturnType<typeof createInitialState>;
+```
+
+状态通常包含：
+- **Region**：用 `createRegion()` 创建的空间区域
+- **parts**：`Record<string, Part>` 游戏棋子集合
+- 游戏特有的字段：当前玩家、分数、回合数等
+
+### 注册命令
+
+使用 `registration.add()` 注册命令。Schema 字符串定义了命令格式：
+
+```ts
+registration.add('play <player> <row:number> <col:number>', async function (cmd) {
+    const [player, row, col] = cmd.params as [PlayerType, number, number];
+
+    // this.context 是 MutableSignal<GameState>
     this.context.produce(state => {
-        state.score[player] += 1;
+        state.parts[piece.id] = piece;
     });
 
+    return { winner: null };
+});
+```
+
+#### Schema 语法
+
+| 语法 | 含义 |
+|---|---|
+| `name` | 命令名 |
+| `<param>` | 必填参数（字符串） |
+| `<param:number>` | 必填参数（自动转为数字） |
+| `[--flag]` | 可选标志 |
+| `[-x:number]` | 可选选项（带类型） |
+
+#### 命令处理器中的 this
+
+命令处理器中的 `this` 是 `CommandRunnerContext<MutableSignal<TState>>`：
+
+```ts
+registration.add('myCommand <arg>', async function (cmd) {
+    // 读取状态
+    const state = this.context.value;
+
+    // 修改状态
+    this.context.produce(draft => {
+        draft.currentPlayer = 'O';
+    });
+
+    // 提示玩家输入
+    const result = await this.prompt(
+        'confirm <action>',
+        (command) => {
+            // 验证函数：返回 null 表示有效，返回 string 表示错误信息
+            return null;
+        },
+        this.context.value.currentPlayer  // currentPlayer 参数可选
+    );
+
+    // 调用子命令
+    const subResult = await this.run<{ score: number }>(`score ${player}`);
+    if (subResult.success) {
+        console.log(subResult.result.score);
+    }
+
+    // 返回命令结果
     return { success: true };
 });
 ```
 
-### Running a Game
+### 使用 prompt 等待玩家输入
+
+`this.prompt()` 是处理玩家输入的核心方法。它会暂停命令执行，等待外部通过 `host.onInput()` 提交输入：
 
 ```ts
-import { createGameContext } from 'boardgame-core';
-import { registry, createInitialState } from './my-game';
+registration.add('turn <player> <turn:number>', async function (cmd) {
+    const [turnPlayer, turnNumber] = cmd.params as [PlayerType, number];
 
-const game = createGameContext(registry, createInitialState);
+    // 等待玩家输入
+    const playCmd = await this.prompt(
+        'play <player> <row:number> <col:number>',  // 期望的输入格式
+        (command) => {
+            const [player, row, col] = command.params as [PlayerType, number, number];
 
-// Run commands through the context
-const result = await game.commands.run('place 2 3');
-if (result.success) {
-    console.log(result.result);
-}
+            // 验证逻辑
+            if (player !== turnPlayer) {
+                return `Invalid player: ${player}`;
+            }
+            if (row < 0 || row > 2 || col < 0 || col > 2) {
+                return `Invalid position: (${row}, ${col})`;
+            }
+            if (isCellOccupied(this.context, row, col)) {
+                return `Cell (${row}, ${col}) is already occupied`;
+            }
 
-// Access reactive game state
-console.log(game.state.value.score.white);
+            return null; // 验证通过
+        },
+        this.context.value.currentPlayer  // 可选：标记当前等待的玩家
+    );
+
+    // 验证通过后，playCmd 是已解析的命令对象
+    const [player, row, col] = playCmd.params as [PlayerType, number, number];
+
+    // 执行放置
+    placePiece(this.context, row, col, player);
+
+    return { winner: checkWinner(this.context) };
+});
 ```
 
-### Handling Player Input
+验证函数返回 `null` 表示输入有效，返回 `string` 表示错误信息。
 
-Commands can prompt for player input using `this.prompt()`. Use `promptQueue.pop()` to wait for prompt events:
+### 使用 setup 命令驱动游戏循环
+
+`setup` 命令通常作为游戏的入口点，负责驱动整个游戏循环：
 
 ```ts
-import { createGameContext } from 'boardgame-core';
-import { registry, createInitialState } from './my-game';
+registration.add('setup', async function () {
+    const { context } = this;
 
-const game = createGameContext(registry, createInitialState);
+    while (true) {
+        const currentPlayer = context.value.currentPlayer;
+        const turnNumber = context.value.turn + 1;
 
-// Start a command that will prompt for input
-const runPromise = game.commands.run('turn X 1');
+        // 运行回合命令
+        const turnOutput = await this.run<{ winner: WinnerType }>(
+            `turn ${currentPlayer} ${turnNumber}`
+        );
+        if (!turnOutput.success) throw new Error(turnOutput.error);
 
-// Wait for the prompt event
-const promptEvent = await game.commands.promptQueue.pop();
-console.log(promptEvent.schema.name); // e.g. 'play'
+        // 更新状态
+        context.produce(state => {
+            state.winner = turnOutput.result.winner;
+            if (!state.winner) {
+                state.currentPlayer = state.currentPlayer === 'X' ? 'O' : 'X';
+                state.turn = turnNumber;
+            }
+        });
 
-// Validate and submit player input
-const error = promptEvent.tryCommit('play X 1 2');
+        // 游戏结束条件
+        if (context.value.winner) break;
+    }
 
-if (error) {
-    console.log('Invalid move:', error);
-    // tryCommit can be called again with corrected input
-} else {
-    // Input accepted, command continues
-    const result = await runPromise;
-}
+    return context.value;
+});
 ```
 
-If the player needs to cancel instead of committing:
+### 使用 Part 和 Region
+
+#### 创建和放置 Part
 
 ```ts
-promptEvent.cancel('player quit');
-```
+import { createPart, createRegion, moveToRegion } from 'boardgame-core';
 
-### Managing Game Lifecycle with GameHost
+// 创建区域
+const board = createRegion('board', [
+    { name: 'row', min: 0, max: 2 },
+    { name: 'col', min: 0, max: 2 },
+]);
 
-For a cleaner game lifecycle (setup, reset, dispose), use `GameHost`:
-
-```ts
-import { createGameHost } from 'boardgame-core';
-import { registry, createInitialState } from './my-game';
-
-// Create a game host with a setup command
-const host = createGameHost(
-    { registry, createInitialState },
-    'setup'  // command to run when setting up/resetting the game
+// 创建棋子
+const piece = createPart<{ owner: string }>(
+    { regionId: 'board', position: [1, 1], owner: 'white' },
+    'piece-1'
 );
 
-// Reactive state — use in effects or computed values
-console.log(host.state.value.currentPlayer);
-console.log(host.status.value); // 'created' | 'running' | 'disposed'
-
-// Check if a prompt is active and what schema it expects
-const schema = host.activePromptSchema.value;
-if (schema) {
-    console.log('Waiting for:', schema.name, schema.params);
-}
-
-// Submit player input to the active prompt
-const error = host.onInput('play X 1 2');
-if (error) {
-    console.log('Invalid move:', error);
-}
-
-// Reset the game (cancels active prompt, resets state, runs setup)
-await host.setup('setup');
-
-// Dispose when done (cleans up listeners, cancels active prompts)
-host.dispose();
+// 放入状态
+state.produce(draft => {
+    draft.parts[piece.id] = piece;
+    draft.board.childIds.push(piece.id);
+    draft.board.partMap['1,1'] = piece.id;
+});
 ```
 
-The `GameHost` provides:
-- **`state`**: `ReadonlySignal<TState>` — reactive game state
-- **`status`**: `ReadonlySignal<'created' | 'running' | 'disposed'>` — lifecycle status
-- **`activePromptSchema`**: `ReadonlySignal<CommandSchema | null>` — reactive current prompt schema
-- **`onInput(input)`**: Submit input to active prompt, returns error or null
-- **`setup(command)`**: Reset and reinitialize the game
-- **`dispose()`**: Clean up all resources
-- **`on(event, listener)`**: Listen to `'setup'` or `'dispose'` events
-
-## Sample Games
-
-### Tic-Tac-Toe
-
-The simplest example. Shows the basic command loop, 2D board regions, and win detection.
-
-See [`src/samples/tic-tac-toe.ts`](src/samples/tic-tac-toe.ts).
-
-### Boop
-
-A more complex game with piece types (kittens/cats), supply management, the "boop" push mechanic, and graduation rules.
-
-See [`src/samples/boop/index.ts`](src/samples/boop/index.ts).
-
-## Region System
+#### 创建 Part 池
 
 ```ts
-import { applyAlign, shuffle, moveToRegion } from 'boardgame-core';
+import { createPartPool } from 'boardgame-core';
 
-// Compact cards in a hand towards the start
-applyAlign(handRegion, parts);
+// 从池中抽取
+const pool = createPartPool<{ type: string }>(
+    { regionId: 'supply', type: 'kitten' },
+    10,
+    'kitten'
+);
 
-// Shuffle positions of all parts in a region
-shuffle(handRegion, parts, rng);
-
-// Move a part from one region to another
-moveToRegion(part, sourceRegion, targetRegion, [0, 0]);
+const piece = pool.draw();       // 取出一个
+pool.return(piece);              // 放回
+pool.remaining();                // 剩余数量
 ```
 
-## Command Parsing
+#### 区域操作
 
 ```ts
-import { parseCommand, parseCommandSchema, validateCommand } from 'boardgame-core';
+import { applyAlign, shuffle, moveToRegion, isCellOccupied } from 'boardgame-core';
 
-// Parse a command string
-const cmd = parseCommand('move card1 hand --force -x 10');
-// { name: 'move', params: ['card1', 'hand'], flags: { force: true }, options: { x: '10' } }
+// 检查格子是否被占用
+if (isCellOccupied(state.parts, 'board', [1, 1])) { ... }
 
-// Define and validate against a schema
-const schema = parseCommandSchema('move <from> <to> [--force] [-x: number]');
-const result = validateCommand(cmd, schema);
-// { valid: true }
+// 对齐排列（紧凑排列）
+applyAlign(handRegion, state.parts);
+
+// 打乱位置
+shuffle(deckRegion, state.parts, rng);
+
+// 移动到其他区域
+moveToRegion(piece, sourceRegion, targetRegion, [0, 0]);
 ```
 
-## Random Number Generation
+### 使用 RNG
 
 ```ts
 import { createRNG } from 'boardgame-core';
 
-const rng = createRNG(12345);
-rng.nextInt(6);       // 0-5
-rng.next();           // [0, 1)
-rng.next(100);        // [0, 100)
-rng.setSeed(999);     // reseed
+const rng = createRNG(12345);  // 种子
+rng.nextInt(6);                // 0-5
+rng.next();                    // [0, 1)
 ```
 
-## API Reference
+### 完整示例：井字棋
 
-### Core
+参考 [`src/samples/tic-tac-toe.ts`](src/samples/tic-tac-toe.ts)，包含：
+- 2D 棋盘区域
+- 玩家轮流输入
+- 胜负判定
+- 完整的游戏循环
 
-| Export | Description |
+---
+
+## API 参考
+
+### 核心
+
+| 导出 | 说明 |
 |---|---|
-| `IGameContext` | Base interface for the game context (state, commands) |
-| `createGameContext(registry, initialState?)` | Create a game context instance. `initialState` can be an object or factory function |
-| `createGameCommandRegistry<TState>()` | Create a command registry with fluent `.add()` API |
-| `GameHost<TState>` | Game lifecycle manager class with setup/reset/dispose |
-| `createGameHost(module, setupCommand, options?)` | Create a GameHost instance from a game module |
-| `GameHostStatus` | Type: `'created' \| 'running' \| 'disposed'` |
-| `GameHostOptions` | Options: `{ autoStart?: boolean }` |
+| `IGameContext` | 游戏上下文基础接口 |
+| `createGameContext(registry, initialState?)` | 创建游戏上下文实例 |
+| `createGameCommandRegistry<TState>()` | 创建命令注册表，返回带 `.add()` 的对象 |
+| `createGameModule(module)` | 辅助函数，标记一个对象为 GameModule |
+| `GameHost<TState>` | 游戏生命周期管理类 |
+| `createGameHost(module, options?)` | 从 GameModule 创建 GameHost |
+| `GameHostStatus` | 类型: `'created' \| 'running' \| 'disposed'` |
 
-### Parts
+### 棋子 (Parts)
 
-| Export | Description |
+| 导出 | 说明 |
 |---|---|
-| `Part<TMeta>` | Type representing a game piece with sides, position, and region. `TMeta` for game-specific fields |
-| `PartTemplate<TMeta>` | Template type for creating parts (excludes `id`, requires metadata) |
-| `PartPool<TMeta>` | Pool of parts with `draw()`, `return()`, and `remaining()` methods. `parts` field is `Record<string, Part>` |
-| `createPart(template, id)` | Create a single part from a template |
-| `createParts(template, count, idPrefix)` | Create multiple identical parts with auto-generated IDs |
-| `createPartPool(template, count, idPrefix)` | Create a pool of parts for lazy loading |
-| `mergePartPools(...pools)` | Merge multiple part pools into one |
-| `findPartById(parts, id)` | Find a part by ID in a Record |
-| `isCellOccupied(parts, regionId, position)` | Check if a cell is occupied |
-| `getPartAtPosition(parts, regionId, position)` | Get the part at a specific position |
-| `flip(part)` | Cycle to the next side |
-| `flipTo(part, side)` | Set to a specific side |
-| `roll(part, rng)` | Randomize side using RNG |
+| `Part<TMeta>` | 游戏棋子类型 |
+| `PartTemplate<TMeta>` | 创建棋子的模板类型 |
+| `PartPool<TMeta>` | 棋子池 |
+| `createPart(template, id)` | 创建单个棋子 |
+| `createParts(template, count, idPrefix)` | 批量创建相同棋子 |
+| `createPartPool(template, count, idPrefix)` | 创建棋子池 |
+| `mergePartPools(...pools)` | 合并多个棋子池 |
+| `findPartById(parts, id)` | 按 ID 查找棋子 |
+| `isCellOccupied(parts, regionId, position)` | 检查格子是否被占用 |
+| `flip(part)` / `flipTo(part, side)` / `roll(part, rng)` | 翻面/随机面 |
 
-### Regions
+### 区域 (Regions)
 
-| Export | Description |
+| 导出 | 说明 |
 |---|---|
-| `Region` | Type for spatial grouping of parts with axis-based positioning |
-| `RegionAxis` | Axis definition with min/max/align |
-| `createRegion(id, axes)` | Create a new region |
-| `applyAlign(region, parts)` | Compact parts according to axis alignment |
-| `shuffle(region, parts, rng)` | Randomize part positions |
-| `moveToRegion(part, sourceRegion?, targetRegion, position?)` | Move a part to another region. `sourceRegion` is optional for first placement |
-| `moveToRegionAll(parts, sourceRegion?, targetRegion, positions?)` | Move multiple parts to another region. `parts` is `Record<string, Part>`. `sourceRegion` is optional for first placement |
-| `removeFromRegion(part, region)` | Remove a part from its region |
+| `Region` / `RegionAxis` | 区域类型 |
+| `createRegion(id, axes)` | 创建区域 |
+| `applyAlign(region, parts)` | 紧凑排列 |
+| `shuffle(region, parts, rng)` | 打乱位置 |
+| `moveToRegion(part, sourceRegion?, targetRegion, position?)` | 移动棋子到其他区域 |
+| `moveToRegionAll(parts, sourceRegion?, targetRegion, positions?)` | 批量移动 |
+| `removeFromRegion(part, region)` | 从区域移除棋子 |
 
-### Commands
+### 命令 (Commands)
 
-| Export | Description |
+| 导出 | 说明 |
 |---|---|
-| `parseCommand(input)` | Parse a command string into a `Command` object |
-| `parseCommandSchema(schema)` | Parse a schema string into a `CommandSchema` |
-| `validateCommand(cmd, schema)` | Validate a command against a schema |
-| `parseCommandWithSchema(cmd, schema)` | Parse and validate in one step |
-| `applyCommandSchema(cmd, schema)` | Apply schema validation and return validated command |
-| `createCommandRegistry<TContext>()` | Create a new command registry |
-| `registerCommand(registry, runner)` | Register a command runner |
-| `unregisterCommand(registry, name)` | Remove a command from the registry |
-| `hasCommand(registry, name)` | Check if a command exists |
-| `getCommand(registry, name)` | Get a command runner by name |
-| `runCommand(registry, context, input)` | Parse and run a command string |
-| `runCommandParsed(registry, context, command)` | Run a pre-parsed command |
-| `createCommandRunnerContext(registry, context)` | Create a command runner context |
-| `PromptEvent` | Event dispatched when a command prompts for input |
-| `CommandRunnerEvents` | Event types: `prompt` (when prompt starts), `promptEnd` (when prompt completes) |
+| `parseCommand(input)` | 解析命令字符串 |
+| `parseCommandSchema(schema)` | 解析 Schema 字符串 |
+| `validateCommand(cmd, schema)` | 验证命令 |
+| `Command` / `CommandSchema` / `CommandResult` | 命令相关类型 |
+| `PromptEvent` | 玩家输入提示事件 |
+| `createRNG(seed?)` | 创建种子 RNG |
 
-### Utilities
+---
 
-| Export | Description |
-|---|---|
-| `createRNG(seed?)` | Create a seeded RNG instance |
-| `Mulberry32RNG` | Mulberry32 PRNG class |
-
-## Scripts
+## 脚本
 
 ```bash
-npm run build       # Build ESM bundle + declarations to dist/
-npm run test        # Run tests in watch mode
-npm run test:run    # Run tests once
-npm run typecheck   # Type check with TypeScript
+npm run build       # 构建 ESM bundle + 类型声明到 dist/
+npm run test        # 以 watch 模式运行测试
+npm run test:run    # 运行测试一次
+npm run typecheck   # TypeScript 类型检查
 ```
 
 ## License