command-guide

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Claude Code Command Selection Guide

Claude Code 命令选择指南

This skill helps you choose the most appropriate command, agent, or skill for different scenarios.

该Skill可帮助你针对不同场景选择最合适的命令、Agent或Skill。

Quick Decision Flowchart

快速决策流程图

mermaid

graph TD
    A[User Request] --> B{Request Type?}
    B -->|New Feature| C[/plan]
    B -->|Bug Fix| D[/tdd or build-error-resolver]
    B -->|Code Review| E[/code-review or code-reviewer agent]
    B -->|Testing| F[/e2e or tdd-guide agent]
    B -->|Context Too Long| G[/compact]
    B -->|Documentation| H[/docs or docs-lookup agent]
    B -->|Looping Task| I[/loop]
    B -->|Security Review| J[security-reviewer agent]

    C --> K[planner agent]
    D --> L{Build Failed?}
    L -->|Yes| M[build-error-resolver]
    L -->|No| N[tdd-guide]
    E --> O[code-reviewer]
    F --> P[e2e-runner]

mermaid

graph TD
    A[User Request] --> B{Request Type?}
    B -->|New Feature| C[/plan]
    B -->|Bug Fix| D[/tdd or build-error-resolver]
    B -->|Code Review| E[/code-review or code-reviewer agent]
    B -->|Testing| F[/e2e or tdd-guide agent]
    B -->|Context Too Long| G[/compact]
    B -->|Documentation| H[/docs or docs-lookup agent]
    B -->|Looping Task| I[/loop]
    B -->|Security Review| J[security-reviewer agent]

    C --> K[planner agent]
    D --> L{Build Failed?}
    L -->|Yes| M[build-error-resolver]
    L -->|No| N[tdd-guide]
    E --> O[code-reviewer]
    F --> P[e2e-runner]

1. Built-in Slash Commands

1. 内置斜杠命令

Session Management Commands

会话管理命令

Command	Use Case	Example
`/compact`	Context too long (>150K tokens), slow response, task phase transition	`/compact` or auto-trigger
`/clear`	Start fresh conversation, clear history	`/clear`
`/loop`	Periodic task execution, automated looping work	`/loop 5m check build status`
`/help`	View help, learn commands	`/help`
`/fast`	Need faster response (Opus 4.6 only)	`/fast`
`/model`	Switch model	`/model sonnet`

命令	适用场景	示例
`/compact`	上下文过长（>150K tokens）、响应缓慢、任务阶段切换	`/compact` 或自动触发
`/clear`	开启全新对话、清除历史记录	`/clear`
`/loop`	周期性任务执行、自动化循环工作	`/loop 5m check build status`
`/help`	查看帮助、了解命令	`/help`
`/fast`	需要更快响应（仅Opus 4.6支持）	`/fast`
`/model`	切换模型	`/model sonnet`

Development Workflow Commands

开发工作流命令

Command	Use Case	Activation Timing
`/plan`	Start new feature, architecture refactor, complex tasks	Enter Plan Mode
`/tdd`	Write tests, TDD development workflow	When test guidance needed
`/e2e`	E2E testing, critical user flow verification	When browser testing needed
`/code-review`	Code quality review	After writing code
`/build-fix`	Build failure, type errors	When build fails
`/learn`	Extract patterns from session, learning	Before session ends
`/skill-create`	Create new skill from git history	When repeating patterns found

命令	适用场景	触发时机
`/plan`	启动新功能、架构重构、复杂任务	进入规划模式
`/tdd`	编写测试、TDD开发工作流	需要测试指导时
`/e2e`	端到端测试、关键用户流程验证	需要浏览器测试时
`/code-review`	代码质量评审	编写代码完成后
`/build-fix`	构建失败、类型错误	构建失败时
`/learn`	从会话中提取模式、学习总结	会话结束前
`/skill-create`	从Git历史创建新Skill	发现重复模式时

Documentation & Query Commands

文档与查询命令

Command	Use Case	Example
`/docs`	Update project documentation	`/docs`
`/update-codemaps`	Update code maps	`/update-codemaps`
`/remember`	Save memory to memory system	`/remember user prefers concise output`
`/tasks`	View task list	`/tasks`

命令	适用场景	示例
`/docs`	更新项目文档	`/docs`
`/update-codemaps`	更新代码映射	`/update-codemaps`
`/remember`	将记忆保存到记忆系统	`/remember user prefers concise output`
`/tasks`	查看任务列表	`/tasks`

2. Agents Selection

2. Agent选择

Development Workflow Agents

开发工作流Agent

Agent	Trigger Condition	Purpose
`planner`	Complex feature request, architectural decision	Create implementation plan
`architect`	System design, tech stack selection	Architecture analysis and decisions
`tdd-guide`	New feature, bug fix	TDD workflow guidance
`code-reviewer`	Invoke immediately after writing code	Code quality review
`security-reviewer`	Handling auth, API, sensitive data	Security vulnerability detection

Agent	触发条件	用途
`planner`	复杂功能需求、架构决策	创建实现方案
`architect`	系统设计、技术栈选择	架构分析与决策
`tdd-guide`	新功能开发、Bug修复	TDD工作流指导
`code-reviewer`	编写代码完成后立即调用	代码质量评审
`security-reviewer`	处理认证、API、敏感数据	安全漏洞检测

Problem Solving Agents

问题解决Agent

Agent	Trigger Condition	Purpose
`build-error-resolver`	Invoke immediately when build fails	Fix build/type errors
`e2e-runner`	Critical user flows, before PR	E2E test execution
`refactor-cleaner`	Code maintenance, dead code cleanup	Dead code detection and cleanup
`doc-updater`	Update docs, codemaps	Documentation sync

Agent	触发条件	用途
`build-error-resolver`	构建失败时立即调用	修复构建/类型错误
`e2e-runner`	关键用户流程、PR提交前	端到端测试执行
`refactor-cleaner`	代码维护、死代码清理	死代码检测与清理
`doc-updater`	更新文档、代码映射	文档同步

Research & Exploration Agents

研究与探索Agent

Agent	Trigger Condition	Purpose
`Explore`	Codebase exploration, file finding	Quick codebase exploration
`general-purpose`	Complex multi-step tasks	General task handling
`docs-lookup`	Query library/framework docs	Get latest API documentation

Agent	触发条件	用途
`Explore`	代码库探索、文件查找	快速代码库探索
`general-purpose`	复杂多步骤任务	通用任务处理
`docs-lookup`	查询库/框架文档	获取最新API文档

3. Skills Selection

3. Skill选择

Workflow Skills

工作流Skill

Skill	Trigger Timing	Purpose
`tdd-workflow`	Developing new feature/fixing bug	Complete TDD workflow guidance
`verification-loop`	After feature completion, before PR	Comprehensive verification (build/test/lint/security)
`strategic-compact`	Long session, context pressure	Guide when to manually `/compact`

Skill	触发时机	用途
`tdd-workflow`	开发新功能/修复Bug	完整TDD工作流指导
`verification-loop`	功能完成后、PR提交前	全面验证（构建/测试/代码检查/安全）
`strategic-compact`	长会话、上下文压力大	指导何时手动调用/compact

Architecture & Pattern Skills

架构与模式Skill

Skill	Trigger Timing	Purpose
`frontend-patterns`	Frontend development	React/Next.js/Vue best practices
`backend-patterns`	Backend development	API/service architecture patterns
`api-design`	API design	RESTful/API design standards
`mcp-server-patterns`	MCP server development	MCP configuration and patterns

Skill	触发时机	用途
`frontend-patterns`	前端开发	React/Next.js/Vue最佳实践
`backend-patterns`	后端开发	API/服务架构模式
`api-design`	API设计	RESTful/API设计标准
`mcp-server-patterns`	MCP服务器开发	MCP配置与模式

Testing Skills

测试Skill

Skill	Trigger Timing	Purpose
`e2e-testing`	E2E testing needs	Playwright test generation
`security-review`	Security review needs	OWASP Top 10 detection

Skill	触发时机	用途
`e2e-testing`	端到端测试需求	Playwright测试生成
`security-review`	安全评审需求	OWASP Top 10检测

Research Skills

研究Skill

Skill	Trigger Timing	Purpose
`deep-research`	Need deep research	Multi-round search and research
`exa-search`	Need web search	Web content search
`documentation-lookup`	Query library docs	Context7 documentation query

Skill	触发时机	用途
`deep-research`	需要深度研究	多轮搜索与研究
`exa-search`	需要网页搜索	网页内容搜索
`documentation-lookup`	查询库文档	Context7文档查询

4. Scenario Decision Matrix

4. 场景决策矩阵

By Task Phase

按任务阶段划分

Phase	Recommended Tool Combination	Reason
Requirements Analysis	`planner` + `Explore`	Plan first, explore later
Architecture Design	`architect` + `api-design` skill	Professional architecture guidance
Pre-Development	`tdd-guide` + `tdd-workflow` skill	Test first
During Development	Direct edit + quick iteration	Stay in flow
Post-Development	`code-reviewer` + `verification-loop`	Quality gate
Testing Phase	`e2e-runner` + `e2e-testing` skill	Complete test coverage
Before PR	`security-reviewer` + `verification-loop`	Final verification
Build Failure	`build-error-resolver`	Focused fix

阶段	推荐工具组合	原因
需求分析	`planner` + `Explore`	先规划，再探索
架构设计	`architect` + `api-design` Skill	专业架构指导
开发前准备	`tdd-guide` + `tdd-workflow` Skill	先测试
开发中	直接编辑 + 快速迭代	保持流畅开发状态
开发后	`code-reviewer` + `verification-loop`	质量把关
测试阶段	`e2e-runner` + `e2e-testing` Skill	完整测试覆盖
PR提交前	`security-reviewer` + `verification-loop`	最终验证
构建失败	`build-error-resolver`	专注修复

By Problem Type

按问题类型划分

Problem	Invoke Immediately	Note
Build failure	`build-error-resolver`	Minimal changes, quick fix
Type error	`build-error-resolver`	TypeScript specialist
Bug fix	`tdd-guide`	Write test then fix
Security vulnerability	`security-reviewer`	OWASP detection
Poor code quality	`code-reviewer`	Immediate review
Missing documentation	`doc-updater`	Auto update
Dead code	`refactor-cleaner`	Safe cleanup

问题	立即调用	说明
构建失败	`build-error-resolver`	最小改动，快速修复
类型错误	`build-error-resolver`	TypeScript专家
Bug修复	`tdd-guide`	先写测试再修复
安全漏洞	`security-reviewer`	OWASP检测
代码质量差	`code-reviewer`	立即评审
文档缺失	`doc-updater`	自动更新
死代码	`refactor-cleaner`	安全清理

By Development Type

按开发类型划分

Development Type	Skills Combination
Frontend feature	`frontend-patterns` + `tdd-workflow`
Backend API	`backend-patterns` + `api-design` + `tdd-workflow`
MCP server	`mcp-server-patterns` + `tdd-workflow`
Database	`database-reviewer` agent
Security feature	`security-reviewer` + `security-review` skill

开发类型	Skill组合
前端功能	`frontend-patterns` + `tdd-workflow`
后端API	`backend-patterns` + `api-design` + `tdd-workflow`
MCP服务器	`mcp-server-patterns` + `tdd-workflow`
数据库	`database-reviewer` Agent
安全功能	`security-reviewer` + `security-review` Skill

5. Parallel Execution Strategy

5. 并行执行策略

Parallelizable Scenarios

可并行场景

Recommended: Launch multiple independent tasks simultaneously

Scenario: Preparing PR after code completion

Agent 1: code-reviewer (code quality)
Agent 2: security-reviewer (security review)
Agent 3: e2e-runner (E2E tests)

Scenario: Large refactor analysis

Agent 1: architect (architecture analysis)
Agent 2: Explore (code exploration)
Agent 3: refactor-cleaner (dead code detection)

Sequential Execution Required

需顺序执行场景

Cannot parallelize: Dependencies exist

Scenario: Fixing build error

Sequence: build-error-resolver -> test verification -> code-reviewer

Scenario: New feature development

Sequence: planner -> tdd-guide (write tests) -> implementation -> code-reviewer

不可并行：存在依赖关系

场景：修复构建错误

顺序：build-error-resolver -> 测试验证 -> code-reviewer

场景：新功能开发

顺序：planner -> tdd-guide（编写测试）-> 实现 -> code-reviewer

6. Auto-Trigger Rules

6. 自动触发规则

Invoke Without User Request

无需用户请求自动调用

Situation	Auto Action
Code written/modified	Immediately invoke `code-reviewer`
Build fails	Immediately invoke `build-error-resolver`
Complex feature request	Immediately invoke `planner`
Handling auth/sensitive data	Immediately invoke `security-reviewer`
New feature/bug fix	Immediately invoke `tdd-guide`
Architectural decision	Immediately invoke `architect`

场景	自动操作
代码编写/修改	立即调用 `code-reviewer`
构建失败	立即调用 `build-error-resolver`
复杂功能需求	立即调用 `planner`
处理认证/敏感数据	立即调用 `security-reviewer`
新功能/Bug修复	立即调用 `tdd-guide`
架构决策	立即调用 `architect`

7. Context Management Timing

7. 上下文管理时机

Indicator	Trigger `/compact`
Token > 150K	Immediately compact
Slow response	Suggest compact
Task phase switch	Compact at boundary
Major milestone completed	Compact then continue
Debugging ends -> new task	Clear debug traces

Best Practices:

Compact after research, before implementation (preserve plan)
Compact after milestone completion (clear intermediate state)
Don't compact mid-implementation (lose variables/paths)

指标	触发/compact
Token数 > 150K	立即压缩
响应缓慢	建议压缩
任务阶段切换	在阶段边界压缩
主要里程碑完成	压缩后继续
调试结束 -> 新任务	清除调试痕迹

最佳实践:

研究完成后、实现前压缩（保留规划内容）
里程碑完成后压缩（清除中间状态）
实现过程中不要压缩（丢失变量/路径）

8. Command Cheat Sheet

8. 命令速查表

Development Workflow:
/plan        -> Enter planning mode (complex tasks)
/tdd         -> TDD workflow
/e2e         -> E2E testing
/code-review -> Code review
/build-fix   -> Fix build

Session Management:
/compact     -> Compact context
/clear       -> Clear session
/loop        -> Looping task
/fast        -> Fast mode

Documentation & Memory:
/docs        -> Update docs
/remember    -> Save memory
/tasks       -> View tasks

Help:
/help        -> View all commands

Development Workflow:
/plan        -> Enter planning mode (complex tasks)
/tdd         -> TDD workflow
/e2e         -> E2E testing
/code-review -> Code review
/build-fix   -> Fix build

Session Management:
/compact     -> Compact context
/clear       -> Clear session
/loop        -> Looping task
/fast        -> Fast mode

Documentation & Memory:
/docs        -> Update docs
/remember    -> Save memory
/tasks       -> View tasks

Help:
/help        -> View all commands

9. Usage Examples

9. 使用示例

Example 1: New Feature Development

示例1：新功能开发

User: Add user authentication feature

Workflow:

/plan -> planner agent creates plan
tdd-guide -> write tests
Implementation -> edit code
code-reviewer -> code review
security-reviewer -> security review (auth sensitive)
e2e-runner -> E2E tests
/compact -> compact after milestone completion

用户：添加用户认证功能

工作流：

/plan -> planner Agent创建方案
tdd-guide -> 编写测试
实现 -> 编辑代码
code-reviewer -> 代码评审
security-reviewer -> 安全评审（认证属于敏感内容）
e2e-runner -> 端到端测试
/compact -> 里程碑完成后压缩上下文

Example 2: Build Failure

示例2：构建失败

User: npm run build failed

Workflow:

build-error-resolver -> analyze error, minimal fix
Verify build success
code-reviewer -> check fix quality

用户：npm run build失败

工作流：

build-error-resolver -> 分析错误，最小化修复
验证构建成功
code-reviewer -> 检查修复质量

Example 3: Code Refactoring

示例3：代码重构

User: Refactor authentication module

Workflow:

architect -> architecture analysis
planner -> implementation plan
refactor-cleaner -> dead code detection
tdd-guide -> ensure test coverage
Implementation -> refactor code
verification-loop -> comprehensive verification

Core Principles:

Plan first, implement later - Use
```
/plan
```
for complex tasks
Test first - Use
```
tdd-guide
```
for new features
Review immediately after coding - Use
```
code-reviewer
```
when code complete
Fix build immediately when failed - Use
```
build-error-resolver
```
Review sensitive code - Use
```
security-reviewer
```
for auth/API
Verify comprehensively before PR - Use
```
verification-loop
```

用户：重构认证模块

工作流：

architect -> 架构分析
planner -> 实现方案
refactor-cleaner -> 死代码检测
tdd-guide -> 确保测试覆盖
实现 -> 重构代码
verification-loop -> 全面验证

核心原则:

先规划，后实现 - 复杂任务使用/plan
先测试 - 新功能使用tdd-guide
编码完成后立即评审 - 代码完成时使用code-reviewer
构建失败立即修复 - 使用build-error-resolver
敏感代码评审 - 认证/API相关使用security-reviewer
PR提交前全面验证 - 使用verification-loop