Initial commit: ClaudeForge v1.0.0

This commit is contained in:
Reza Rezvani
2025-11-12 11:19:48 +01:00
commit 37422c1667
42 changed files with 11812 additions and 0 deletions
+548
View File
@@ -0,0 +1,548 @@
# Architecture Overview
Technical architecture and design decisions for ClaudeForge.
---
## System Design
### Component Architecture
```
┌─────────────────────────────────────────────────────────┐
│ User Project │
│ │
│ User runs: /enhance-claude-md │
└────────────────────┬────────────────────────────────────┘
┌─────────────────────────────────────────────────────────┐
│ Slash Command (enhance-claude-md.md) │
│ │
│ Phase 1: Discovery - Check CLAUDE.md existence │
│ Phase 2: Analysis - Determine initialize/enhance │
│ Phase 3: Task - Invoke skill or agent │
└────────────────────┬────────────────────────────────────┘
┌───────────┴───────────┐
│ │
↓ ↓
┌──────────────────┐ ┌──────────────────┐
│ Guardian Agent │ │ Direct Skill │
│ (Background) │ │ Invocation │
│ │ │ │
│ • SessionStart │ │ • User request │
│ • Manual invoke │ │ • One-time gen │
└────────┬─────────┘ └─────────┬────────┘
│ │
└────────────┬───────────┘
┌────────────────────────┐
│ Skill: claudeforge │
│ │
│ Python Modules: │
│ • workflow.py │
│ • analyzer.py │
│ • validator.py │
│ • template_selector │
│ • generator.py │
└────────────┬───────────┘
┌────────────────────────┐
│ CLAUDE.md Output │
│ │
│ • Root file │
│ • Context files │
│ • Native format │
└────────────────────────┘
```
---
## Data Flow
### 1. New Project Initialization
```
1. User → /enhance-claude-md
2. Command checks: CLAUDE.md not found
3. Command → Skill (initialize mode)
4. Skill workflow.py:
- generate_exploration_prompt()
- Claude explores repository
- analyze_discoveries(exploration_results)
- Returns project_context
5. Skill template_selector.py:
- select_template(project_context)
- Returns template configuration
6. Skill generator.py:
- generate_root_file(template)
- Returns CLAUDE.md content
7. Skill validator.py:
- validate_all(generated_content)
- Returns validation report
8. Output → CLAUDE.md file(s) created
```
### 2. Existing Project Enhancement
```
1. User → /enhance-claude-md
2. Command checks: CLAUDE.md exists
3. Command → Skill (enhance mode)
4. Skill analyzer.py:
- analyze_file(current_content)
- calculate_quality_score()
- Returns analysis report
5. Skill validator.py:
- validate_all(current_content)
- Returns validation results
6. Skill shows user:
- Quality score (0-100)
- Missing sections
- Recommendations
7. User confirms enhancement
8. Skill generator.py:
- merge_with_existing(content, sections)
- Returns enhanced content
9. Output → CLAUDE.md updated
```
### 3. Background Maintenance
```
1. SessionStart event
2. Guardian agent triggered
3. Agent checks git changes:
- git diff --name-status HEAD~10
- git diff package.json requirements.txt
4. Agent determines significance:
- Files changed > 5?
- New dependencies?
- New directories?
5. If significant:
- Agent → Skill (enhance mode)
- Skill updates specific sections
- Agent validates changes
6. Output → CLAUDE.md synced
```
---
## Module Design
### workflow.py
**Purpose:** Interactive initialization for new projects
**Key Classes:**
- `InitializationWorkflow` - Main orchestrator
**Key Methods:**
```python
check_claude_md_exists() bool
generate_exploration_prompt() str
analyze_discoveries(results: Dict) Dict[str, Any]
_detect_project_type(results) str
_detect_tech_stack(results) List[str]
_estimate_team_size(results) str
_detect_development_phase(results) str
_detect_workflows(results) List[str]
_should_use_modular(results) bool
```
**Detection Logic:**
- Project type: Check for frontend/, backend/, src/ patterns
- Tech stack: Parse package.json, requirements.txt, go.mod
- Team size: Analyze git contributors, project complexity
- Phase: Check for CI/CD, production configs
- Workflows: Detect test/, .github/, documentation patterns
---
### analyzer.py
**Purpose:** Analyze existing CLAUDE.md files
**Key Classes:**
- `CLAUDEMDAnalyzer` - File analyzer
**Quality Scoring Algorithm:**
```python
def calculate_quality_score(self) int:
score = 0
# Length appropriateness (25 points)
if 20 <= line_count <= 300:
score += 25
elif 300 < line_count <= 400:
score += 15 # Warn: consider modular
else:
score += 5 # Poor: too short or too long
# Section completeness (25 points)
required = ["Core Principles", "Tech Stack", "Workflow"]
found = len([s for s in required if s in sections])
score += (found / len(required)) * 25
# Formatting quality (20 points)
# Check: headings, code blocks, lists, links
score += formatting_score
# Content specificity (15 points)
# Check: project-specific vs. generic
score += specificity_score
# Modular organization (15 points)
# Check: context files if needed
score += modular_score
return min(score, 100)
```
---
### validator.py
**Purpose:** Validate against best practices
**Validation Categories:**
1. **Length Validation**
- Recommended: 20-300 lines
- Warning: 300-400 lines (suggest modular)
- Error: < 20 or > 400 lines
2. **Structure Validation**
- Required sections: Core Principles, Tech Stack, Workflow
- Recommended sections: Testing, Error Handling
- Heading hierarchy: H1 → H2 → H3 (no skips)
3. **Formatting Validation**
- Balanced code blocks (open/close)
- Valid markdown syntax
- No broken links
4. **Completeness Validation**
- Has code examples
- Lists tech stack with versions
- Includes setup instructions
5. **Anti-Pattern Detection**
- No hardcoded secrets (API keys, tokens)
- No TODO placeholders
- No broken reference links
---
### template_selector.py
**Purpose:** Select appropriate template
**Selection Matrix:**
| Project Type | Team Size | Lines | Template |
|--------------|-----------|-------|----------|
| CLI/Library | Solo | 50-75 | minimal |
| Web App | Small | 100-150 | core |
| API | Small | 125-175 | api-focused |
| Full-Stack | Medium | 200-300 | detailed |
| Enterprise | Large | Modular | root + contexts |
**Modular Recommendation Logic:**
```python
def recommend_modular_structure(context: Dict) bool:
return (
context['type'] == 'fullstack' or
context['team_size'] in ['medium', 'large'] or
context['phase'] in ['production', 'enterprise'] or
len(context['tech_stack']) > 5
)
```
---
### generator.py
**Purpose:** Generate CLAUDE.md content
**Generation Modes:**
1. **Root File (Navigation Hub)**
- Quick Navigation section
- Core Principles (high-level)
- Tech Stack summary
- Links to context files
2. **Context File (Specific Area)**
- Detailed guidelines for backend/, frontend/, etc.
- Tech-specific patterns
- Common commands for that area
3. **Section Generation (Individual)**
- Generate single section
- Merge with existing content
**Native Format Template:**
```markdown
# CLAUDE.md
[Overview paragraph]
## Project Structure
```
project/
├── src/
│ ├── components/
│ └── services/
└── tests/
```
## File Structure
- `src/` - Source code
- `tests/` - Test files
## Setup & Installation
```bash
npm install
npm run dev
```
## Architecture
[Key design decisions]
## Core Principles
1. Principle one
2. Principle two
## Tech Stack
- React 18
- TypeScript 5
- Node.js 20
## Common Commands
```bash
npm run build # Build project
npm test # Run tests
```
```
---
## Integration Points
### Skill ↔ Slash Command
**File:** `command/enhance-claude-md.md`
```yaml
# YAML frontmatter
allowed-tools: Bash, Read, Glob, Skill
# Phase 3: Task section
I can invoke the `claude-md-enhancer` skill...
```
Claude Code recognizes skill name and loads Python modules.
### Skill ↔ Guardian Agent
**File:** `agent/claude-md-guardian.md`
```yaml
# YAML frontmatter
tools: Bash, Read, Write, Edit, Grep, Glob, Skill
model: haiku
# Agent workflow section
I invoke the `claude-md-enhancer` skill...
```
Agent uses `haiku` model for token efficiency, invokes skill for updates.
### Agent ↔ Git
Agent detects changes via bash commands:
```bash
git diff --name-status HEAD~10
git log --since="1 week ago" --oneline
git diff HEAD~10 -- package.json requirements.txt
```
Triggers update if:
- 5+ files changed
- Dependencies modified
- New directories created
---
## Design Decisions
### Why Python for Skill Modules?
- **Portability:** Standard library only, no dependencies
- **Readability:** Clear logic for community contributions
- **Performance:** Adequate for file analysis/generation
- **Integration:** Claude Code supports Python natively
### Why Separate Slash Command and Agent?
- **Slash Command:** User-initiated, interactive, immediate feedback
- **Agent:** Background, automatic, non-intrusive
- **Flexibility:** User chooses explicit control vs. automation
### Why Quality Scoring Algorithm?
- **Objectivity:** Consistent evaluation across projects
- **Actionable:** Specific areas to improve
- **Educational:** Users learn best practices
- **Gamification:** Encourages quality improvement
### Why Modular Architecture Support?
- **Scalability:** Large projects need organization
- **Context:** Different areas have different needs
- **Maintainability:** Easier to update specific sections
- **Team:** Different teams own different files
---
## Performance Considerations
### Token Efficiency
**Guardian Agent uses `haiku` model:**
- Routine updates: ~500-1000 tokens
- Targeted section updates only
- Saves 70-80% tokens vs. full regeneration
**Slash Command uses default model (sonnet):**
- Interactive, user-facing
- Requires better understanding
- More complex reasoning
### File Size Limits
- Single file: Max 400 lines (prefer 300)
- Modular files: Each 150-300 lines
- Total project: Unlimited with modular
### Caching Strategy
No caching implemented (stateless design):
- Each invocation reads fresh files
- Ensures accuracy with latest changes
- Simple implementation
---
## Security Considerations
### Anti-Pattern Detection
**validator.py** checks for:
- Hardcoded secrets: `API_KEY=`, `password=`, `token=`
- Placeholder TODOs: `TODO`, `FIXME`, `XXX`
- Broken links: Invalid URL patterns
### File Permissions
Installation respects user permissions:
- User-level: `~/.claude/` (user writable)
- Project-level: `./.claude/` (project writable)
- No system-level changes
### Git Integration
Agent only reads git:
- `git diff` (read-only)
- `git log` (read-only)
- `git status` (read-only)
- No git write operations
---
## Extensibility
### Adding New Project Types
1. Update `workflow.py``_detect_project_type()`
2. Add detection patterns
3. Update `template_selector.py` → selection matrix
4. Create new template in `skill/examples/`
### Adding New Tech Stacks
1. Update `workflow.py``_detect_tech_stack()`
2. Add file detection (e.g., `Cargo.toml` for Rust)
3. Update `generator.py` → tech-specific sections
### Adding New Validation Rules
1. Update `validator.py``validate_all()`
2. Add new validation method
3. Return validation result in standard format
---
## Testing Strategy
### Manual Testing
See [QUICK_START.md](QUICK_START.md) for test scenarios.
### Integration Testing
Test entire flow:
1. Install components
2. Run slash command
3. Verify output quality
4. Test guardian agent
5. Validate native format
### Unit Testing
(Not implemented in v1.0.0, planned for v1.1.0)
- Test each Python module independently
- Mock file I/O
- Validate scoring algorithms
---
## Future Architecture
### Planned for v1.1.0
- **VS Code Extension:** Inline editing, real-time validation
- **GitHub Action:** Auto-generate on repo creation
- **Custom Templates:** User-defined template system
- **Analytics:** Usage patterns, effectiveness metrics
### Planned for v2.0.0
- **AI-Powered Suggestions:** Context-aware recommendations
- **Multi-Language Support:** i18n for generated content
- **Web Dashboard:** Project-wide management
- **Plugin System:** Third-party extensions
---
## Related Documentation
- **Implementation Details:** [CLAUDE.md](../CLAUDE.md)
- **Installation:** [INSTALLATION.md](INSTALLATION.md)
- **Quick Start:** [QUICK_START.md](QUICK_START.md)
- **Troubleshooting:** [TROUBLESHOOTING.md](TROUBLESHOOTING.md)
---
**Questions about architecture?** Open an issue: https://github.com/alirezarezvani/ClaudeForge/issues
+490
View File
@@ -0,0 +1,490 @@
# Contributing to ClaudeForge
Thank you for your interest in contributing to ClaudeForge! This guide will help you get started.
---
## Code of Conduct
By participating in this project, you agree to abide by our [Code of Conduct](../.github/CODE_OF_CONDUCT.md).
---
## How Can I Contribute?
### 1. Reporting Bugs
**Before submitting a bug report:**
- Check existing issues to avoid duplicates
- Collect information about the bug
- Test with the latest version
**Submit a bug report:**
- Use the bug report template
- Provide clear title and description
- Include steps to reproduce
- Add relevant logs or screenshots
- Mention your environment (OS, Claude Code version)
**Template:** https://github.com/alirezarezvani/ClaudeForge/issues/new?template=bug_report.md
---
### 2. Suggesting Enhancements
**Before submitting:**
- Check if the feature already exists
- Review planned features in CHANGELOG.md
- Consider if it fits project scope
**Submit an enhancement:**
- Use the feature request template
- Explain the use case
- Describe proposed solution
- Consider alternatives
**Template:** https://github.com/alirezarezvani/ClaudeForge/issues/new?template=feature_request.md
---
### 3. Contributing Code
#### Quick Start
```bash
# Fork the repository
# Clone your fork
git clone https://github.com/YOUR-USERNAME/ClaudeForge.git
cd ClaudeForge
# Create a branch
git checkout -b feature/amazing-feature
# Make your changes
# Test your changes
./install.sh # Test installation
/enhance-claude-md # Test in Claude Code
# Commit your changes
git commit -m "Add amazing feature"
# Push to your fork
git push origin feature/amazing-feature
# Open a Pull Request
```
#### Development Setup
```bash
# Install development dependencies
# (None required - Python standard library only)
# Test Python modules
python3 -c "from skill.analyzer import CLAUDEMDAnalyzer; print('OK')"
# Test installation scripts
./install.sh
# Choose option 2 (project-level) for testing
```
---
## Development Guidelines
### Python Code Style
**Follow PEP 8:**
```python
# Good
def analyze_file(self, content: str) -> Dict[str, Any]:
"""Analyze CLAUDE.md file content."""
pass
# Bad
def analyzeFile(self,content):
pass
```
**Type Hints:**
```python
# Always use type hints
def calculate_score(sections: List[str]) int:
return len(sections) * 10
```
**Docstrings:**
```python
def validate_length(self, content: str) -> bool:
"""
Validate CLAUDE.md file length.
Args:
content: File content as string
Returns:
True if length is valid, False otherwise
"""
pass
```
---
### Markdown Style
**Headings:**
```markdown
# H1 - Document title
## H2 - Main sections
### H3 - Subsections
```
**Code Blocks:**
```markdown
```bash
# Use language-specific syntax highlighting
command here
```
```
**Links:**
```markdown
# Prefer relative links for internal docs
[Architecture](ARCHITECTURE.md)
# Use absolute URLs for external
[GitHub](https://github.com/alirezarezvani/ClaudeForge)
```
---
### Testing Your Changes
#### Test Python Modules
```bash
# Test analyzer
python3 << EOF
from skill.analyzer import CLAUDEMDAnalyzer
content = open('test-CLAUDE.md').read()
analyzer = CLAUDEMDAnalyzer(content)
report = analyzer.analyze_file()
print(f"Quality Score: {report['quality_score']}")
EOF
```
#### Test Installation
```bash
# Test install.sh
./install.sh
# Choose option 2 (project-level)
# Verify all components copied correctly
# Test uninstall
rm -rf ./.claude
```
#### Test in Claude Code
```bash
# Install your changes
./install.sh
# Restart Claude Code
# Test slash command
/enhance-claude-md
# Verify output matches expected behavior
```
---
## Contribution Areas
### 1. Adding New Templates
**Location:** `skill/examples/`
**Steps:**
1. Create new template file (e.g., `rust-cli-CLAUDE.md`)
2. Follow native format structure
3. Update `skill/examples/README.md`
4. Update `template_selector.py` detection logic
5. Add test case to `sample_input.json`
**Example:**
```bash
# Create Rust template
vim skill/examples/rust-cli-CLAUDE.md
# Update selector
vim skill/template_selector.py
# Add: if 'Cargo.toml' in files: return 'rust-cli'
# Test
# Create test Rust project, run /enhance-claude-md
```
---
### 2. Improving Detection Logic
**Location:** `skill/workflow.py`
**Methods to enhance:**
- `_detect_project_type()`
- `_detect_tech_stack()`
- `_estimate_team_size()`
- `_detect_development_phase()`
**Example:**
```python
# Add Flutter detection
def _detect_project_type(self, results):
files = results.get('files', [])
# Add Flutter check
if 'pubspec.yaml' in files:
return 'mobile'
# ... existing logic
```
---
### 3. Enhancing Quality Scoring
**Location:** `skill/analyzer.py`
**Method:** `calculate_quality_score()`
**Ideas:**
- Add more granular checks
- Weight sections by importance
- Detect project-specific customization
**Example:**
```python
def calculate_quality_score(self) -> int:
score = 0
# Add new check: Has examples
if self._has_code_examples():
score += 5 # Bonus points
# ... existing logic
return min(score, 100)
```
---
### 4. Adding New Validation Rules
**Location:** `skill/validator.py`
**Steps:**
1. Add new validation method
2. Call from `validate_all()`
3. Return standard validation result
**Example:**
```python
def validate_examples(self) -> Dict[str, Any]:
"""Validate code examples are present."""
code_blocks = re.findall(r'```.*?```', self.content, re.DOTALL)
return {
'passed': len(code_blocks) >= 3,
'message': 'At least 3 code examples required',
'severity': 'low'
}
```
---
### 5. Documentation Improvements
**Areas:**
- Fix typos or unclear explanations
- Add missing examples
- Improve installation instructions
- Translate to other languages (future)
**Process:**
1. Edit markdown files in `docs/`
2. Test markdown rendering locally
3. Submit pull request
---
## Pull Request Process
### 1. Before Submitting
- [ ] Code follows project style guidelines
- [ ] All tests pass (manual testing required)
- [ ] Documentation updated if needed
- [ ] CHANGELOG.md updated with your changes
- [ ] Commit messages are clear and descriptive
### 2. Submitting
1. **Create Pull Request** from your fork
2. **Fill out PR template** completely
3. **Link related issues** (if any)
4. **Request review** from maintainers
5. **Respond to feedback** promptly
### 3. After Submission
- **CI checks** will run automatically
- **Maintainers** will review your code
- **Address feedback** by pushing new commits
- **Don't force-push** after review started
- **Be patient** - reviews may take a few days
### 4. Merging
Once approved:
- Maintainer will merge your PR
- Your contribution will be in the next release
- You'll be added to contributors list
---
## Commit Message Guidelines
### Format
```
type(scope): Short description
Longer description if needed.
Fixes #123
```
### Types
- `feat`: New feature
- `fix`: Bug fix
- `docs`: Documentation changes
- `style`: Code style changes (no logic change)
- `refactor`: Code refactoring
- `test`: Adding tests
- `chore`: Maintenance tasks
### Examples
```
feat(analyzer): Add code example detection
Adds check for code examples in CLAUDE.md files.
Contributes to quality score calculation.
Fixes #45
```
```
fix(installer): Resolve Windows path issues
Fixes path separator issues on Windows platforms.
Updates install.ps1 to use platform-specific paths.
Fixes #67
```
```
docs(quick-start): Clarify installation steps
Adds more detailed explanations for each step.
Includes troubleshooting for common issues.
```
---
## Release Process
(For maintainers)
### Version Numbering
Follow Semantic Versioning (MAJOR.MINOR.PATCH):
- MAJOR: Breaking changes
- MINOR: New features (backward compatible)
- PATCH: Bug fixes (backward compatible)
### Release Checklist
1. **Update version:**
- `CHANGELOG.md` - Add new version section
- `README.md` - Update version badge
- `skill/SKILL.md` - Update version footer
2. **Create release:**
```bash
git tag -a v1.1.0 -m "Release v1.1.0"
git push origin v1.1.0
```
3. **GitHub Release:**
- Create release from tag
- Copy CHANGELOG excerpt to description
- Attach any binaries if needed
4. **Announce:**
- GitHub Discussions
- Social media (if applicable)
- Update documentation site
---
## Community
### Getting Help
- **GitHub Discussions:** Ask questions, share ideas
- **GitHub Issues:** Report bugs, request features
- **Documentation:** Comprehensive guides in `docs/`
### Staying Updated
- **Watch Repository:** Get notifications for releases
- **Star Repository:** Show support, stay informed
- **Follow Updates:** Check CHANGELOG.md for changes
---
## Recognition
Contributors are recognized in:
- **README.md** - Contributors section
- **CHANGELOG.md** - Version-specific contributors
- **GitHub** - Automatic contributors graph
Every contribution matters, from code to documentation to bug reports!
---
## Questions?
- **General Questions:** GitHub Discussions
- **Bug Reports:** GitHub Issues
- **Security Issues:** Email maintainers directly (see SECURITY.md if created)
- **Feature Requests:** GitHub Issues with feature template
---
## License
By contributing, you agree that your contributions will be licensed under the MIT License.
See [LICENSE](../LICENSE) for details.
---
**Thank you for contributing to ClaudeForge!** 🎉
Your contributions help make CLAUDE.md management better for everyone in the Claude Code community.
+412
View File
@@ -0,0 +1,412 @@
# Installation Guide
Complete installation instructions for ClaudeForge on all supported platforms.
---
## Prerequisites
### Required
- **Claude Code** 2.0 or later
- **Operating System:** macOS, Linux, or Windows
### Recommended
- **Git** (for change detection by guardian agent)
- Terminal access (macOS/Linux) or PowerShell (Windows)
---
## Installation Methods
### Method 1: One-Line Install (Recommended)
#### macOS / Linux
```bash
curl -fsSL https://raw.githubusercontent.com/alirezarezvani/ClaudeForge/main/install.sh | bash
```
#### Windows (PowerShell as Administrator)
```powershell
iwr https://raw.githubusercontent.com/alirezarezvani/ClaudeForge/main/install.ps1 -useb | iex
```
**What this does:**
1. Downloads the installer script
2. Prompts for installation scope (user-level or project-level)
3. Copies all components to appropriate directories
4. Creates backups of any existing installations
5. Optionally installs quality hooks
---
### Method 2: Manual Git Clone
```bash
# Clone the repository
git clone https://github.com/alirezarezvani/ClaudeForge.git
# Navigate to directory
cd ClaudeForge
# Run installer
./install.sh # macOS/Linux
# or
.\install.ps1 # Windows
```
---
### Method 3: Manual Installation (Advanced)
If you prefer to manually copy files:
#### Step 1: Download Components
Download the latest release from:
https://github.com/alirezarezvani/ClaudeForge/releases
#### Step 2: Choose Installation Scope
**User-Level (Recommended)** - Available in all Claude Code projects:
```bash
# macOS/Linux
INSTALL_DIR="$HOME/.claude"
# Windows
$INSTALL_DIR="$env:USERPROFILE\.claude"
```
**Project-Level** - Available only in current project:
```bash
# macOS/Linux
INSTALL_DIR="./.claude"
# Windows
$INSTALL_DIR=".\.claude"
```
#### Step 3: Copy Components
```bash
# Create directories
mkdir -p "$INSTALL_DIR/skills"
mkdir -p "$INSTALL_DIR/commands"
mkdir -p "$INSTALL_DIR/agents"
# Copy skill
cp -r skill "$INSTALL_DIR/skills/claudeforge-skill"
# Copy slash command
cp -r command "$INSTALL_DIR/commands/enhance-claude-md"
# Copy guardian agent
cp agent/claude-md-guardian.md "$INSTALL_DIR/agents/"
```
#### Step 4: Restart Claude Code
Close and restart Claude Code to load the new components.
---
## Installation Scopes
### User-Level Installation
**Location:** `~/.claude/` (or `%USERPROFILE%\.claude` on Windows)
**Advantages:**
- ✅ Available in all Claude Code projects
- ✅ Install once, use everywhere
- ✅ Automatic updates apply globally
**Use When:**
- You work on multiple projects
- You want consistent CLAUDE.md standards across all projects
- You're the primary developer on your machine
### Project-Level Installation
**Location:** `./.claude/` (in your project root)
**Advantages:**
- ✅ Project-specific configuration
- ✅ Version controlled with project (can commit to git)
- ✅ Team members get same tools
**Use When:**
- Different projects need different versions
- You want to version control the tools
- Team collaboration requires shared tooling
---
## Verification
After installation, verify all components are installed correctly:
### Check Skill Installation
```bash
# macOS/Linux
ls -la ~/.claude/skills/claudeforge-skill/
# Windows
dir $env:USERPROFILE\.claude\skills\claudeforge-skill\
```
**Expected output:**
```
analyzer.py
generator.py
template_selector.py
validator.py
workflow.py
SKILL.md
README.md
examples/
```
### Check Command Installation
```bash
# macOS/Linux
ls -la ~/.claude/commands/enhance-claude-md/
# Windows
dir $env:USERPROFILE\.claude\commands\enhance-claude-md\
```
**Expected output:**
```
enhance-claude-md.md
README.md
```
### Check Agent Installation
```bash
# macOS/Linux
ls -la ~/.claude/agents/claude-md-guardian.md
# Windows
dir $env:USERPROFILE\.claude\agents\claude-md-guardian.md
```
### Test in Claude Code
1. Restart Claude Code
2. Open any project
3. Run the command:
```
/enhance-claude-md
```
4. You should see the multi-phase workflow start
---
## Quality Hooks (Optional)
During installation, you'll be asked if you want to install quality hooks.
**What are Quality Hooks?**
- Pre-commit validation that checks CLAUDE.md quality before commits
- Ensures best practices compliance
- Only available for project-level installations
**To Install Hooks:**
```bash
# During installer
# When prompted: "Would you like to install quality hooks?"
# Type: y
# Or manually:
mkdir -p .claude/hooks
cp hooks/pre-commit.sh .claude/hooks/
chmod +x .claude/hooks/pre-commit.sh
```
**To Use Hooks:**
Configure git to use the hook:
```bash
# Add to .git/config or use git config
git config core.hooksPath .claude/hooks
```
---
## Troubleshooting Installation
### Issue: "Installation files not found"
**Cause:** Running installer from wrong directory
**Solution:**
```bash
cd ClaudeForge # Navigate to repository root
./install.sh # Run from correct directory
```
---
### Issue: "Permission denied"
**Cause:** Installer doesn't have execute permission
**Solution:**
```bash
chmod +x install.sh # macOS/Linux
# or run with bash explicitly
bash install.sh
```
---
### Issue: "~/.claude directory not found"
**Cause:** Claude Code not installed or hasn't been run yet
**Solution:**
1. Ensure Claude Code is installed
2. Run Claude Code at least once to create directory structure
3. Or let installer create directories (it will prompt)
---
### Issue: Command not recognized after installation
**Cause:** Claude Code hasn't reloaded components
**Solution:**
1. Fully quit Claude Code (not just close window)
2. Restart Claude Code
3. Wait a few seconds for components to load
4. Try command again: `/enhance-claude-md`
---
### Issue: "Skill not found" error
**Cause:** Skill directory name mismatch
**Solution:**
```bash
# Verify skill directory name is exactly:
ls ~/.claude/skills/
# Should show: claudeforge-skill
# If different, rename:
mv ~/.claude/skills/old-name ~/.claude/skills/claudeforge-skill
```
---
### Issue: Windows installer fails with execution policy error
**Cause:** PowerShell execution policy restricts scripts
**Solution:**
```powershell
# Run PowerShell as Administrator
# Set execution policy temporarily:
Set-ExecutionPolicy -Scope Process -ExecutionPolicy Bypass
# Then run installer:
.\install.ps1
```
---
## Updating ClaudeForge
To update to a newer version:
### Method 1: Using Installer
```bash
# Pull latest changes
cd ClaudeForge
git pull origin main
# Run installer (existing installation will be backed up)
./install.sh
```
### Method 2: Manual Update
```bash
# Backup current installation
mv ~/.claude/skills/claudeforge-skill ~/.claude/skills/claudeforge-skill.backup
# Copy new version
cp -r skill ~/.claude/skills/claudeforge-skill
# Repeat for command and agent
```
---
## Uninstallation
To completely remove ClaudeForge:
### User-Level Uninstall
```bash
# macOS/Linux
rm -rf ~/.claude/skills/claudeforge-skill
rm -rf ~/.claude/commands/enhance-claude-md
rm -f ~/.claude/agents/claude-md-guardian.md
# Windows (PowerShell)
Remove-Item -Recurse -Force $env:USERPROFILE\.claude\skills\claudeforge-skill
Remove-Item -Recurse -Force $env:USERPROFILE\.claude\commands\enhance-claude-md
Remove-Item -Force $env:USERPROFILE\.claude\agents\claude-md-guardian.md
```
### Project-Level Uninstall
```bash
# macOS/Linux
rm -rf ./.claude/skills/claudeforge-skill
rm -rf ./.claude/commands/enhance-claude-md
rm -f ./.claude/agents/claude-md-guardian.md
rm -rf ./.claude/hooks # If quality hooks were installed
# Windows (PowerShell)
Remove-Item -Recurse -Force .\.claude\skills\claudeforge-skill
Remove-Item -Recurse -Force .\.claude\commands\enhance-claude-md
Remove-Item -Force .\.claude\agents\claude-md-guardian.md
Remove-Item -Recurse -Force .\.claude\hooks
```
After uninstalling, restart Claude Code.
---
## Next Steps
After successful installation:
1. **Read Quick Start Guide:** [QUICK_START.md](QUICK_START.md)
2. **Test the command:** Run `/enhance-claude-md` in a project
3. **Review Architecture:** [ARCHITECTURE.md](ARCHITECTURE.md)
4. **Explore Examples:** Check `examples/` directory
---
## Support
If you encounter installation issues not covered here:
- **GitHub Issues:** https://github.com/alirezarezvani/ClaudeForge/issues
- **Troubleshooting Guide:** [TROUBLESHOOTING.md](TROUBLESHOOTING.md)
- **Discussions:** https://github.com/alirezarezvani/ClaudeForge/discussions
---
**Installation successful?** Proceed to [Quick Start Guide](QUICK_START.md) →
+407
View File
@@ -0,0 +1,407 @@
# Quick Start Guide
Get started with ClaudeForge in 5 minutes. This guide walks you through your first CLAUDE.md creation.
---
## 🚀 5-Minute Tutorial
### Step 1: Install ClaudeForge (1 minute)
```bash
# macOS / Linux
curl -fsSL https://raw.githubusercontent.com/alirezarezvani/ClaudeForge/main/install.sh | bash
# Windows (PowerShell)
iwr https://raw.githubusercontent.com/alirezarezvani/ClaudeForge/main/install.ps1 -useb | iex
```
**Choose:** Option 1 (user-level) for global availability
**Restart Claude Code** after installation
---
### Step 2: Navigate to Your Project (30 seconds)
```bash
cd /path/to/your/project
```
**Requirements:**
- Any code project (doesn't need to be git-initialized, but recommended)
- Works with TypeScript, Python, Go, Java, Ruby, and more
---
### Step 3: Run the Command (10 seconds)
In Claude Code, type:
```bash
/enhance-claude-md
```
---
### Step 4: Interactive Workflow (3 minutes)
ClaudeForge will now:
#### Phase 1: Discovery ✓
Claude explores your project:
```
Checking for existing CLAUDE.md... Not found
Exploring repository structure...
Analyzing package.json, requirements.txt, etc...
```
#### Phase 2: Analysis ✓
Claude shows discoveries:
```
Based on my exploration, here's what I discovered:
📦 Project Type: Full-Stack Application
🛠️ Tech Stack: TypeScript, React, Node.js, PostgreSQL, Docker
👥 Team Size: Small (5 developers detected from git history)
🚀 Development Phase: MVP
⚙️ Workflows: TDD, CI/CD
📋 Recommended Structure:
- Root CLAUDE.md (navigation hub, ~100 lines)
- backend/CLAUDE.md (API guidelines, ~150 lines)
- frontend/CLAUDE.md (React guidelines, ~175 lines)
```
#### Phase 3: Confirmation ✓
Claude asks for your approval:
```
Would you like me to create these files with these settings?
```
**You respond:** `Yes` or `Yes, proceed`
#### Phase 4: Generation ✓
Claude creates your CLAUDE.md files:
```
Creating customized CLAUDE.md files...
✅ Created CLAUDE.md (108 lines)
- Quick Navigation section
- 5 Core Principles
- Tech Stack summary
- Common commands
✅ Created backend/CLAUDE.md (156 lines)
- API design guidelines
- Database operations
- Testing requirements
✅ Created frontend/CLAUDE.md (182 lines)
- Component standards
- State management
- Performance tips
```
---
### Step 5: Review Your CLAUDE.md (1 minute)
Open the generated file:
```bash
cat CLAUDE.md
```
**You'll see:**
- **Overview** - Project description
- **Project Structure** - ASCII tree diagram
- **Setup & Installation** - How to get started
- **Architecture** - Key patterns and decisions
- **Core Principles** - Development guidelines
- **Tech Stack** - Technologies used
- **Common Commands** - Build, test, lint
- **Development Workflow** - How to contribute
All in **100% native Claude Code format**!
---
## 🎯 What Just Happened?
1. **Explored** your codebase automatically
2. **Detected** project type, tech stack, team size
3. **Generated** customized CLAUDE.md with:
- Native format compliance
- Tech-specific best practices
- Team-appropriate complexity
4. **Created** modular files (if needed)
**Result:** Future Claude Code sessions now have perfect context about your project!
---
## 🔄 Common Workflows
### Scenario A: New Project
```bash
# You: Create a new project
mkdir my-new-app && cd my-new-app
npm init -y
# You: Run ClaudeForge
/enhance-claude-md
# Claude: Explores → Analyzes → Creates CLAUDE.md
```
**Time:** ~2 minutes
---
### Scenario B: Existing Project (No CLAUDE.md)
```bash
# You: Navigate to existing project
cd my-existing-project
# You: Run ClaudeForge
/enhance-claude-md
# Claude: Explores → Detects tech → Creates CLAUDE.md
```
**Time:** ~3 minutes
---
### Scenario C: Improve Existing CLAUDE.md
```bash
# You: Have basic CLAUDE.md
cat CLAUDE.md
# Output: Basic file with just tech stack
# You: Run ClaudeForge
/enhance-claude-md
# Claude: Analyzes current file
# Shows: Quality Score: 55/100
# Missing: Project Structure, Setup, Architecture
# Claude: Adds missing sections
# Result: Quality Score: 55 → 88
```
**Time:** ~2 minutes
---
### Scenario D: Automatic Maintenance
```bash
# You: Make significant changes
npm install react-query
mkdir src/components/auth
# You: Start new Claude Code session
# Guardian agent automatically checks changes
# Agent output:
# ✅ CLAUDE.md updated:
# - Tech Stack: Added react-query
# - Project Structure: Updated diagram
# Changes: 2 sections, 8 lines
```
**Time:** Automatic, ~10 seconds
---
## 📚 Next Steps
### Learn More
- **Architecture:** [ARCHITECTURE.md](ARCHITECTURE.md) - How components work together
- **Examples:** [../examples/](../examples/) - Real-world usage patterns
- **Troubleshooting:** [TROUBLESHOOTING.md](TROUBLESHOOTING.md) - Common issues
### Customize Your CLAUDE.md
After generation, you can:
1. **Edit Core Principles** - Add team-specific guidelines
2. **Add Custom Sections** - Project-specific conventions
3. **Update Tech Stack** - Add missing dependencies
4. **Refine Workflows** - Adjust to your process
### Enable Automatic Maintenance
Set up the guardian agent to keep CLAUDE.md current:
```bash
# Agent runs automatically at session start
# Or invoke manually after major changes:
Claude, invoke claude-md-guardian to update CLAUDE.md
```
---
## 💡 Pro Tips
### Tip 1: Use Modular Architecture for Large Projects
If your CLAUDE.md exceeds 300 lines, split it:
```bash
/enhance-claude-md
# Tell Claude: "Use modular architecture"
# Result: Separate files for backend/, frontend/, database/
```
### Tip 2: Regenerate When Tech Stack Changes
```bash
# You: Added new major framework
npm install @nestjs/core
# You: Regenerate with new context
/enhance-claude-md
# Claude: Updates Tech Stack and patterns
```
### Tip 3: Validate Before Committing
```bash
# You: Made manual edits to CLAUDE.md
# You: Want to check quality
/enhance-claude-md
# Claude: Analyzes and shows quality score
# If score < 80, Claude suggests improvements
```
### Tip 4: Team Consistency
Install at user-level on all team machines:
```bash
# Each team member runs:
curl -fsSL https://raw.githubusercontent.com/alirezarezvani/ClaudeForge/main/install.sh | bash
# Choose option 1 (user-level)
# Result: Consistent CLAUDE.md standards across team
```
---
## 🎓 Understanding the Output
### What's in a Generated CLAUDE.md?
**Native Format Sections (Always Included):**
1. **Overview** - What this project does
2. **Project Structure** - ASCII tree diagram
3. **File Structure** - Directory explanations
4. **Setup & Installation** - Getting started steps
5. **Architecture** - Key design decisions (for complex projects)
**Team-Specific Sections:**
6. **Core Principles** - Development guidelines
7. **Tech Stack** - Technologies with versions
8. **Development Workflow** - How to contribute
9. **Testing Requirements** - Test standards
10. **Common Commands** - Build, test, lint scripts
**Optional Sections (Based on Project Type):**
- **Error Handling** - Error patterns
- **Performance Guidelines** - Optimization tips
- **Security Checklist** - Security requirements
- **Deployment Process** - How to deploy
---
## ⚡ Quick Reference
### Commands
| Command | Purpose |
|---------|---------|
| `/enhance-claude-md` | Initialize or enhance CLAUDE.md |
| `Claude, invoke claude-md-guardian` | Manually trigger maintenance |
### Quality Scores
| Score | Meaning | Action |
|-------|---------|--------|
| 0-40 | Poor | Major improvements needed |
| 41-70 | Fair | Add missing sections |
| 71-85 | Good | Minor improvements |
| 86-100 | Excellent | Maintain current quality |
### File Size Guidelines
| Team Size | Recommended Lines | Structure |
|-----------|-------------------|-----------|
| Solo | 50-100 | Single file |
| Small (2-9) | 100-200 | Single file |
| Medium (10-50) | 200-300 | Single or modular |
| Large (50+) | Modular | Multiple context files |
---
## 🆘 Having Issues?
### Issue: Command not recognized
**Solution:** Restart Claude Code after installation
### Issue: Quality score too low
**Solution:** Run `/enhance-claude-md` and let Claude add missing sections
### Issue: Generated file too generic
**Solution:** Provide more context:
```
/enhance-claude-md
"This is a Python FastAPI microservice project with PostgreSQL"
```
### More Help
See [TROUBLESHOOTING.md](TROUBLESHOOTING.md) for detailed solutions.
---
## 🎉 Success!
You now have a fully functional CLAUDE.md file that:
- ✅ Follows native Claude Code format
- ✅ Includes project-specific context
- ✅ Provides clear development guidelines
- ✅ Automatically maintains itself (via guardian agent)
**Future Claude Code sessions will be significantly more productive!**
---
## 📖 Further Reading
- **Installation Guide:** [INSTALLATION.md](INSTALLATION.md)
- **Architecture Deep Dive:** [ARCHITECTURE.md](ARCHITECTURE.md)
- **Contributing:** [CONTRIBUTING.md](CONTRIBUTING.md)
- **GitHub Repository:** https://github.com/alirezarezvani/ClaudeForge
---
**Ready to create amazing projects with ClaudeForge? Let's build! 🚀**
+551
View File
@@ -0,0 +1,551 @@
# Troubleshooting Guide
Common issues and solutions for ClaudeForge.
---
## Installation Issues
### Issue: Command not recognized after installation
**Symptoms:**
- `/enhance-claude-md` shows "command not found"
- Slash command doesn't autocomplete
**Causes:**
- Claude Code hasn't reloaded components
- Incorrect installation path
- Insufficient permissions
**Solutions:**
1. **Restart Claude Code completely:**
```bash
# Quit Claude Code entirely (not just close window)
# Reopen Claude Code
# Wait 5-10 seconds for components to load
```
2. **Verify installation paths:**
```bash
# macOS/Linux
ls -la ~/.claude/commands/enhance-claude-md/
ls -la ~/.claude/skills/claudeforge-skill/
ls -la ~/.claude/agents/claude-md-guardian.md
# Windows
dir %USERPROFILE%\.claude\commands\enhance-claude-md\
dir %USERPROFILE%\.claude\skills\claudeforge-skill\
dir %USERPROFILE%\.claude\agents\claude-md-guardian.md
```
3. **Reinstall with correct permissions:**
```bash
chmod +x install.sh # macOS/Linux
./install.sh
```
---
### Issue: "Skill not found" error
**Symptoms:**
- Error message: "claude-md-enhancer skill not found"
- Slash command runs but fails
**Cause:** Skill directory name mismatch or missing skill files
**Solutions:**
1. **Check skill directory name (must be exact):**
```bash
ls ~/.claude/skills/
# Must show: claudeforge-skill
```
2. **Verify skill files exist:**
```bash
ls ~/.claude/skills/claudeforge-skill/
# Must contain: SKILL.md, analyzer.py, generator.py, etc.
```
3. **Reinstall skill only:**
```bash
rm -rf ~/.claude/skills/claudeforge-skill
cp -r skill ~/.claude/skills/claudeforge-skill
```
---
### Issue: Windows PowerShell execution policy error
**Symptoms:**
- Error: "execution of scripts is disabled on this system"
**Cause:** PowerShell security settings
**Solution:**
```powershell
# Run PowerShell as Administrator
Set-ExecutionPolicy -Scope Process -ExecutionPolicy Bypass
.\install.ps1
```
---
## Usage Issues
### Issue: Quality score unexpectedly low
**Symptoms:**
- Quality score < 50
- Many "missing section" warnings
**Causes:**
- Basic CLAUDE.md without native format sections
- File too short (< 20 lines) or too long (> 400 lines)
- Missing required sections
**Solutions:**
1. **Let ClaudeForge add missing sections:**
```bash
/enhance-claude-md
# Claude will identify missing sections
# Confirm enhancement
# Quality score will improve
```
2. **Check file length:**
```bash
wc -l CLAUDE.md
# Recommended: 20-300 lines
# If > 300: Consider modular architecture
```
3. **Manually add required sections:**
- Core Principles
- Tech Stack
- Workflow Instructions
- Project Structure (ASCII diagram)
- Setup & Installation
---
### Issue: Generated content too generic
**Symptoms:**
- CLAUDE.md doesn't mention specific tech stack
- No project-specific guidelines
- Feels like template without customization
**Causes:**
- Project detection failed (no package.json, requirements.txt, etc.)
- Insufficient project context
**Solutions:**
1. **Provide explicit context:**
```bash
/enhance-claude-md
# When Claude asks, specify:
"This is a Python FastAPI microservice with PostgreSQL, Redis, and Docker. Team of 8 developers, MVP phase."
```
2. **Add project files before running:**
```bash
# Ensure detection files exist:
touch package.json # Node.js
touch requirements.txt # Python
touch go.mod # Go
touch Cargo.toml # Rust
```
3. **Manual customization after generation:**
- Edit generated CLAUDE.md
- Add project-specific conventions
- Include team standards
---
### Issue: Modular architecture not recommended
**Symptoms:**
- Single CLAUDE.md generated for large project
- File exceeds 300 lines
**Causes:**
- Project type not detected as `fullstack`
- Team size estimated as `small`
**Solutions:**
1. **Explicitly request modular:**
```bash
/enhance-claude-md
# Tell Claude:
"Use modular architecture with separate files for backend, frontend, and database"
```
2. **Create backend/ and frontend/ directories:**
```bash
mkdir -p backend frontend
# Then run /enhance-claude-md
```
3. **Manual split if needed:**
```bash
# Create context files manually:
touch backend/CLAUDE.md
touch frontend/CLAUDE.md
# Then run enhancement on each:
cd backend && /enhance-claude-md
cd frontend && /enhance-claude-md
```
---
### Issue: Guardian agent not updating automatically
**Symptoms:**
- Made significant changes
- Started new session
- CLAUDE.md not updated
**Causes:**
- Agent not installed
- Changes below significance threshold
- Git repository not initialized
**Solutions:**
1. **Verify agent installation:**
```bash
ls ~/.claude/agents/claude-md-guardian.md
```
2. **Check git repository:**
```bash
git status
# Agent requires git for change detection
```
3. **Manually invoke agent:**
```bash
# In Claude Code:
Claude, invoke claude-md-guardian to update CLAUDE.md
```
4. **Lower significance threshold (if needed):**
Edit `agent/claude-md-guardian.md`:
```markdown
# Change from 5 to 3 files minimum
- Less than 5 files modified # Change to 3
```
---
## Output Issues
### Issue: CLAUDE.md missing Project Structure diagram
**Symptoms:**
- No ASCII tree diagram in output
- Native format incomplete
**Cause:** Template generation skipped structure section
**Solution:**
1. **Regenerate with explicit request:**
```bash
/enhance-claude-md
# Tell Claude:
"Add Project Structure section with ASCII tree diagram"
```
2. **Manual addition:**
```markdown
## Project Structure
```
project/
├── src/
│ ├── components/
│ └── services/
└── tests/
```
```
---
### Issue: Setup & Installation section generic
**Symptoms:**
- Says "npm install" for Python project
- Incorrect commands
**Cause:** Tech stack detection mismatch
**Solution:**
1. **Verify tech stack detection:**
```bash
# Check project files:
ls package.json requirements.txt go.mod
```
2. **Regenerate with correct context:**
```bash
/enhance-claude-md
"This is a Python project, not Node.js"
```
3. **Manual fix:**
Edit CLAUDE.md and correct commands
---
## Performance Issues
### Issue: Guardian agent too slow
**Symptoms:**
- Agent takes > 30 seconds
- Multiple updates in single session
**Causes:**
- Model set to sonnet instead of haiku
- Full regeneration instead of targeted update
**Solutions:**
1. **Verify agent model:**
```bash
grep "model:" ~/.claude/agents/claude-md-guardian.md
# Should be: model: haiku
```
2. **Check agent workflow:**
- Should use targeted section updates
- Not full file regeneration
---
### Issue: Slash command timeout
**Symptoms:**
- Command starts but never completes
- "Timeout" error
**Causes:**
- Large repository exploration
- Network issues (if fetching remote content)
**Solutions:**
1. **Run in smaller scope:**
```bash
cd specific-directory
/enhance-claude-md
```
2. **Skip exploration, provide context directly:**
```bash
/enhance-claude-md
"Skip exploration. Create CLAUDE.md for TypeScript React app with PostgreSQL."
```
---
## Integration Issues
### Issue: Git commands fail in agent
**Symptoms:**
- Agent error: "git: command not found"
- Agent can't detect changes
**Cause:** Git not installed or not in PATH
**Solutions:**
1. **Install git:**
```bash
# macOS
brew install git
# Ubuntu/Debian
sudo apt-get install git
# Windows
# Download from git-scm.com
```
2. **Verify git installation:**
```bash
git --version
# Should show: git version 2.x
```
---
### Issue: Quality hooks not working
**Symptoms:**
- Pre-commit hook doesn't run
- No validation before commit
**Causes:**
- Hook not executable
- Git not configured to use hooks
- Wrong hooks directory
**Solutions:**
1. **Make hook executable:**
```bash
chmod +x .claude/hooks/pre-commit.sh
```
2. **Configure git:**
```bash
git config core.hooksPath .claude/hooks
```
3. **Test hook manually:**
```bash
./.claude/hooks/pre-commit.sh
# Should validate CLAUDE.md
```
---
## Validation Errors
### Issue: "File length exceeds recommended maximum"
**Symptoms:**
- Validation warning: File > 300 lines
- Quality score penalty
**Cause:** Single file too large
**Solution:**
1. **Split into modular architecture:**
```bash
/enhance-claude-md
"Convert to modular architecture with separate backend and frontend files"
```
2. **Remove unnecessary content:**
- Delete redundant sections
- Move detailed examples to separate docs
- Link to external documentation
---
### Issue: "Missing required sections"
**Symptoms:**
- Validation fails
- Lists missing: Core Principles, Tech Stack, etc.
**Solution:**
1. **Let ClaudeForge add them:**
```bash
/enhance-claude-md
# Claude will identify and add missing sections
```
2. **Manual addition:**
Add required sections following native format
---
## Common Error Messages
### "CLAUDE.md not found"
**Meaning:** File doesn't exist (expected for new projects)
**Action:** Continue with initialization workflow
---
### "Quality score: 0/100"
**Meaning:** Empty or invalid file
**Action:** Regenerate completely
---
### "Invalid project structure"
**Meaning:** Cannot detect project type
**Action:** Add project files (package.json, etc.) or provide context manually
---
### "Template selection failed"
**Meaning:** No matching template for project
**Action:** Provide more specific project context
---
## Getting Help
If your issue isn't covered here:
1. **Check GitHub Issues:**
https://github.com/alirezarezvani/ClaudeForge/issues
2. **Search Discussions:**
https://github.com/alirezarezvani/ClaudeForge/discussions
3. **Open New Issue:**
- Describe the problem
- Include steps to reproduce
- Attach relevant logs or errors
- Mention your OS and Claude Code version
4. **Review Documentation:**
- [INSTALLATION.md](INSTALLATION.md)
- [QUICK_START.md](QUICK_START.md)
- [ARCHITECTURE.md](ARCHITECTURE.md)
- [CLAUDE.md](../CLAUDE.md)
---
## Debug Mode
To get more detailed output:
```bash
# Run installer with verbose output:
bash -x install.sh
# Check Claude Code logs:
# macOS: ~/Library/Logs/Claude Code/
# Linux: ~/.config/claude-code/logs/
# Windows: %APPDATA%\Claude Code\logs\
```
---
**Still having issues?** We're here to help!
Open an issue: https://github.com/alirezarezvani/ClaudeForge/issues/new