claude-loop

claude-loop

Pure Claude CLI autonomy - iteratively fixes your repo until everything works. No hardcoded patterns, just raw AI power.

What It Does

claude-loop gives Claude complete freedom to:

• 🔄 Iteratively debug your entire repository
• 🧪 Test every button, form, and feature
• 🔧 Fix broken functionality automatically
• 🚀 Deploy Task agents for parallel debugging
• ✨ Simplify complexity by leveraging AI instead of hardcoded patterns

Installation

npm install -g claude-loop

Usage

# Just run it - auto-installs MCPs and starts debugging
claude-loop

# Or use the loop command with options
claude-loop loop --max-iterations 20

Features

🔌 Auto-installs MCPs

On first run, automatically detects and installs:

• VUDA - Visual UI Debug Agent for testing UI
• BrowserMCP - Browser automation capabilities
• Sequential Thinking - Step-by-step problem solving

📊 Enhanced Progress Tracking

🔄 Iteration 3/10
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Progress: ████████████░░░░░░░░░░░░░░░░░░ 30%
Elapsed: 2m 15s
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📡 Continuing conversation with Claude...
Focus: Form validation and submission

🤖 Full Claude Autonomy

• Uses --dangerously-skip-permissions for complete freedom
• All tools enabled: Bash, Read, Write, Edit, Task, etc.
• Creates parallel Task agents for complex debugging

🎯 Smart Focus Areas

Each iteration focuses on specific aspects:

• Button functionality and click handlers
• Form validation and submission
• Navigation and routing
• Data persistence and API calls
• Error handling and edge cases
• UI consistency and responsiveness
• Performance and optimization
• Security and input validation
• Accessibility and user experience
• Final cleanup and polish

How It Works

• Analyzes your repository - Runs tree, grep, and other commands to understand structure
• Deploys Task agents - Creates parallel agents to debug different areas
• Fixes issues iteratively - Each iteration focuses on specific problems
• Continues until done - Uses claude -c to maintain context across iterations

Example Output

$ claude-loop

🔄 Claude Loop - Real Iterative Debugging with Claude CLI

🔧 Checking MCP installations...

✓ VUDA (Visual UI Debug Agent) available
✓ Browser MCP available

📡 Starting Claude session:
Command: claude
Max turns: 30
Tools: Bash, Read, Write, Edit, MultiEdit, Grep, Glob, LS, WebFetch, WebSearch, NotebookRead, NotebookEdit, Task
Mode: Full autonomy (--dangerously-skip-permissions)

[Claude begins debugging your repository...]

Requirements

• Claude CLI installed (pip install claude)
• Node.js >= 14.0.0
• Git repository to debug

What Gets Fixed

• 🔘 Broken buttons without click handlers
• 📝 Forms that don't validate or submit
• 🔗 Dead links and 404 routes
• 💾 Data that doesn't persist
• 🎨 UI inconsistencies
• 🐛 Unhandled errors
• 📱 Responsive issues
• ♿ Accessibility problems
• 🔧 Unnecessary complexity

License

MIT