Claude Reconstruction
Claude Code Engineering Configuration System - Make every session efficient, stable, and reproducible
English | 简体中文
What is This?
Claude Reconstruction is a battle-tested Claude Code Engineering Configuration System that includes:
- Error Knowledge Base - 10+ common error patterns and prevention measures
- Decision Tree - Tool selection guide for 50+ scenarios
- Workflows - Standardized task execution processes
- Capability Documentation - Complete MCP/Skills/Plugins reference
- Best Practices - Coding standards and methodologies
Why Do You Need It?
| Problem | Solution |
|---|---|
| Same errors keep occurring | Error knowledge base + self-check checklists |
| Don't know which tool to use | Decision tree for quick navigation |
| Low task execution efficiency | Standardized workflows |
| Knowledge loss between sessions | Persistent configuration system |
System Requirements
| Requirement | Version/Description |
|---|---|
| Claude Code | >= 1.0.0 |
| Operating System | macOS / Linux / Windows |
| Shell | Bash (Unix/Linux/macOS) / PowerShell (Windows) |
| Node.js | >= 14.0.0 (Optional, for npm installation) |
Quick Start
Installation
Method 1: Clone and Install
# Clone repository
git clone https://github.com/Arxchibobo/claude-Reconstruction.git
cd claude-Reconstruction
# Unix/Linux/macOS
chmod +x scripts/install.sh
./scripts/install.sh
# Windows PowerShell
.\scripts\install.ps1
Method 2: Manual Installation
- Download the repository
- Copy files from
core/directory to~/.claude/ - Copy other directories (errors, capabilities, etc.) to
~/.claude/
Verify Installation
Run verification script:
# Unix/Linux/macOS
./scripts/verify.sh
# Windows PowerShell
.\scripts\verify.ps1
Or start Claude Code, and you should see:
- High-frequency error reminders
- Quick decision tree
- Work mode confirmation
Directory Structure
claude-reconstruction/
├── README.md # This file
├── core/ # Core configuration
│ ├── CLAUDE.md # Main configuration
│ ├── QUICK_START.md # Session startup checklist
│ └── DECISION_TREE.md # Capability decision tree
├── errors/ # Error knowledge base
│ ├── ERROR_CATALOG.md # Error catalog
│ ├── system-errors/ # System-level errors
│ └── project-errors/ # Project-level errors
├── capabilities/ # Capability documentation
│ ├── mcp-servers.md # Complete MCP guide
│ ├── skills-guide.md # Skills usage guide
│ └── plugins-auto.md # Plugins auto-activation
├── workflows/ # Workflows
│ ├── auto-execution.md # Auto-execution mode
│ └── data-analysis.md # Data analysis workflow
├── learning/ # Learning resources
│ └── AI_WORKFLOW_INSIGHTS.md
├── references/ # References
│ └── BEST_PRACTICES.md # Best practices
├── automation/ # Automation configuration
│ └── hooks.md # Hooks configuration
├── delegator/ # Delegation system
│ └── README.md # Delegator documentation
└── scripts/ # Installation scripts
├── install.sh # Unix installation script
└── install.ps1 # Windows installation script
Core Features
1. Error Knowledge Base
10 high-frequency errors with prevention measures:
| ID | Error | Self-Check Question |
|---|---|---|
| E001 | Async without parallelization | Using Promise.all()? |
| E002 | Polling without timeout | Set maxAttempts? |
| E003 | Error not re-thrown | Throw in catch block? |
| E004 | SQL without CTE | Pre-filter data? |
| ... | ... | ... |
2. Decision Tree
Need external data? → MCP (database/observability/chart)
Need automation? → Skills (/commit, /write-tests)
Need suggestions? → Plugins (auto-activated)
3. Work Mode
Plan → Confirm → Execute → Verify
4 Critical Blockers (Only situations where asking is allowed):
- Missing critical credentials
- Multiple conflicting approaches
- Contradictory requirements
- Irreversible high-risk operations
4. Capability Layers
| Layer | Tool | Purpose |
|---|---|---|
| Layer 1 | MCP Servers | External data access |
| Layer 2 | Skills | Automated tasks |
| Layer 3 | Plugins | Expert advice (auto-activated) |
Usage Examples
Data Analysis
User: Analyze user growth last month
Claude:
1. Query user data with database MCP
2. Process data locally
3. Generate trend chart with chart MCP
4. Output analysis report
Feature Development
User: Add user registration feature
Claude:
1. Create TodoList plan
2. Show plan and wait for confirmation
3. Execute completely (no questions)
4. Generate acceptance report
Git Operations
User: /commit
Claude:
1. Run git status to see changes
2. Analyze change content
3. Generate commit message
4. Wait for confirmation then commit
Custom Configuration
Add Project-Specific Errors
Create a new file in errors/project-errors/:
# my-project-errors.md
## E101: Project-Specific Error
**Description**: Error description
**Self-Check**: Self-check questions
**Solution**: Code examples
Add Custom Skill
Create in ~/.claude/commands/:
# my-skill.md
> Describe skill purpose
## Execution Steps
1. Step one
2. Step two
Configure Hooks
Add to ~/.claude/settings.json:
{
"hooks": {
"SessionStart": [
{
"matcher": "",
"hooks": [
{ "type": "command", "command": "cat ~/.claude/startup.md" }
]
}
]
}
}
Contributing
Contributions are welcome! Please:
- Fork this repository
- Create a feature branch
- Submit your changes
- Create a Pull Request
Contribution Areas
- New error patterns
- Workflow optimization
- Documentation improvements
- Bug fixes
See CONTRIBUTING.md for detailed guidelines.
License
MIT License - See LICENSE for details
Acknowledgments
Thanks to all developers who contribute to the Claude Code ecosystem.
Happy Coding with Claude! 🚀