Skip to content

CLAUDE.md

Latest commit

 

History

History
203 lines (162 loc) · 7.12 KB

CLAUDE.md

File metadata and controls

203 lines (162 loc) · 7.12 KB

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

🏗️ Repository Architecture

This is a Claude Code multi-agent workflow system that transforms manual development commands into automated expert team workflows. It implements a Requirements-Driven Development approach with 90% quality gates.

Core System Components

1. Commands Directory (/commands/)

11 specialized slash commands for targeted development tasks:

  • /requirements-pilot - Complete automated workflow (requirements → code → review → test)
  • /ask - Architecture consultation without code changes
  • /code - Feature implementation with constraints
  • /debug - UltraThink systematic problem analysis
  • /test - Comprehensive testing strategy
  • /review - Multi-dimensional code validation
  • /optimize - Performance optimization coordination
  • /refactor - Quality-gated code refactoring
  • /bugfix - Systematic bug resolution workflow
  • /docs - Documentation generation
  • /think - Advanced analysis and strategic thinking

2. Agents Directory (/agents/)

Two-tier agent architecture:

Tier 1: Contains-Studio Agents (/agents/) - 40+ specialized agents organized by:

  • Engineering: ai-engineer, backend-architect, devops-automator, frontend-developer, mobile-app-builder, rapid-prototyper, test-writer-fixer
  • Design: brand-guardian, ui-designer, ux-researcher, visual-storyteller, whimsy-injector
  • Marketing: app-store-optimizer, content-creator, growth-hacker, tiktok-strategist
  • Product: feedback-synthesizer, sprint-prioritizer, trend-researcher
  • Project Management: experiment-tracker, project-shipper, studio-producer
  • Studio Operations: analytics-reporter, finance-tracker, infrastructure-maintainer

Tier 2: Requirements-Driven Agents (/agents/zhisheng/) - 9 specialized agents:

  • requirements-generate - Create implementation-ready technical specs
  • requirements-code - Direct implementation with minimal architecture overhead
  • requirements-review - Practical code review focused on functionality
  • requirements-testing - Functional validation and integration testing
  • bugfix - Defect analysis and resolution
  • bugfix-verify - Objective fix validation
  • code - Development coordination
  • debug - Systematic debugging with UltraThink
  • optimize - Performance optimization coordination

🎯 Key Workflows

Requirements-Driven Full Automation

/requirements-pilot "企业用户管理系统，JWT认证，RBAC权限，500并发用户"

Automatically executes:

  1. Phase 1: Interactive requirements confirmation (90+ quality gate)
  2. Approval Gate: Explicit user approval required
  3. Phase 2: Automated sub-agent chain execution
    • requirements-generate → requirements-code → requirements-review → testing decision

Manual Expert Orchestration

/ask "微服务架构设计考虑"
/code "实现JWT认证中间件"
test "创建并发测试套件"
review "验证安全最佳实践"

📊 Quality Gate System

Requirements Quality (100-point system):

  • Functional Clarity (30 pts) - Clear specs, user interactions, success criteria
  • Technical Specificity (25 pts) - Integration points, constraints, performance needs
  • Implementation Completeness (25 pts) - Edge cases, error handling, validation
  • Business Context (20 pts) - User value, priority definition

Code Quality (90% threshold):

  • Functional correctness
  • Integration quality
  • Code maintainability
  • Performance adequacy

🔧 Usage Patterns

Complex Feature Development

# Full automation with quality gates
/requirements-pilot "OAuth2认证系统，支持刷新令牌，集成企业LDAP，刷新令牌轮换"

# Manual step-by-step for learning/debugging
/ask "OAuth2架构最佳实践"
/code "实现核心认证逻辑"
/test "创建安全测试套件"
review "验证实现质量"

Bug Resolution

# Systematic bug fixing
/bugfix "用户登录在高并发时偶尔失败"

# Manual debugging
/debug "分析登录并发问题"
code "实现修复方案"
bugfix-verify "验证修复效果"

Performance Optimization

optimize "数据库查询性能问题"

🌍 Language Support

  • Chinese: All commands support Chinese input/output
  • Mixed Language: Seamless Chinese/English switching
  • Context-Aware: Automatically detects and responds in appropriate language

📁 Key File Structures

claude_code_settings/
├── commands/          # 11 specialized slash commands
├── agents/           # Multi-tier agent architecture
├── README-zh.md      # Comprehensive Chinese documentation
├── README.md         # English overview
└── CLAUDE.md         # This guidance file

🚀 Quick Start Commands

For New Projects

/requirements-pilot "描述你的第一个功能需求"
# 自动执行完整工作流

For Existing Codebases

/ask "分析现有架构并提出改进建议"
/bugfix "修复已知的生产问题"
optimize "提升关键路径性能"

For Learning/Exploration

/think "分析这个技术决策的利弊"
/ask "解释微服务架构的核心概念"

⚡ Advanced Features

Testing Control Options

  • --test / 要测试 - Force testing execution
  • --no-test / 不要测试 - Skip testing phase
  • Interactive mode - Smart recommendations based on task complexity

Quality Customization

  • Default: 90% quality threshold
  • Override: Set custom thresholds per command
  • Iterative: Automatic optimization loops until threshold met

Context Isolation

  • Each sub-agent operates in independent context
  • No cross-contamination between specialized domains
  • Maintains focus and expertise depth

🎭 Pro Agent Triggers

Some agents activate automatically:

  • studio-coach - Complex multi-agent coordination
  • test-writer-fixer - After code changes/bug fixes
  • whimsy-injector - After UI/UX improvements
  • experiment-tracker - When feature flags added

📈 Development Philosophy

Requirements-Driven: Specifications → Implementation → Validation Quality-Gated: Objective scoring prevents quality degradation Expert Specialization: Domain experts vs. generalists Rapid Iteration: 6-day sprint compatibility Production Ready: Emphasis on working solutions over perfect architecture

🛠️ Common Commands Summary

Command Purpose Automation Level Quality Gate
/requirements-pilot Full development workflow 95% automated 90% threshold
/ask Architecture consultation Manual N/A
/code Feature implementation Semi-automated Review required
/debug Systematic problem analysis Guided Validation
/test Testing strategy Automated Coverage metrics
/optimize Performance tuning Semi-automated Benchmark validation

🔍 Diagnostic Commands

# Check system health
/deploy-check

# Security analysis
/security

# Workflow optimization
/commands/think "分析当前开发流程瓶颈"