.png)
4 months ago
5
Claude Parallel Work is a Model Context Protocol (MCP) server that provides intelligent task breakdown and parallel execution for development projects. It uses containerized isolation and automated dependency analysis to optimize complex workflows.
- Parallel task execution - Break down complex projects into concurrent tasks
- Docker containerization - Isolated execution environment for security
- Intelligent task breakdown - Automated dependency analysis and optimization
- Git-based workflow - Change tracking with diff-based approval process
- Server persistence - SQLite-backed state management with automatic recovery
-
MCP Server Management in Containers - Enable execution containers to run and manage their own MCP servers for enhanced tool capabilities
- Support for dynamic MCP server installation in containers
- Automatic server lifecycle management
- Pre-configured MCP server bundles for common use cases
-
Live Dashboard Updates - Real-time dashboard with enhanced monitoring capabilities
- Live code diff preview with syntax highlighting
- Real-time container resource usage graphs
- Interactive terminal with direct container access for intervening in workers
- Notifications for task status changes
-
Improved Developer Experience
- VS Code extension for direct integration
- Task execution history and analytics
- Performance profiling and optimization suggestions
git clone https://github.com/ddfourtwo/claude-parallel-work
cd claude-parallel-work
chmod +x install.sh
./install.sh
- Use the installed custom slash commands to generate a tasks.json file: /break-down-to-work-plan and /work-plan-from-doc
- Use the installed custom slash commands to kick off orchestration: /orchestrate-tasks
The server provides MCP tools for task management:
- get_next_tasks - Get next tasks to work on
- task_worker - Execute development tasks in containers
- review_changes - Inspect modifications before applying
- apply_changes - Apply approved changes to workspace
- Node.js v20+
- Docker (Docker Desktop or Engine)
- Git repository (working directories must be git repos)
- MCP Client (Claude Desktop, Cursor, Windsurf, etc.)
The monitoring interface provides real-time visibility into task execution:
parallel-work dashboard # Launch dashboard
parallel-work status # Check service status
parallel-work logs -f # View server logs
parallel-work follow # Monitor task execution
Features:
- Repository and task status monitoring
- Container log streaming with terminal interface
- Change diff management and preview
- Live progress updates via WebSocket
# Service management
parallel-work status # Check if services are running
parallel-work restart # Restart crashed server
parallel-work dashboard # Launch monitoring interface
# Monitoring & debugging
parallel-work logs # View recent logs
parallel-work logs -f # Follow logs in real-time
parallel-work follow # Follow task execution
parallel-work health # Run health check
# Configuration
parallel-work config # Show current config
parallel-work version # Show version info
# Build and test
npm run build
npm test
# Development mode
npm run dev
# Dashboard development
npm run dashboard # Start dashboard dev server
npm run start:dashboard # Start enhanced server
- CLI Usage Guide - Complete CLI reference
- Known Issues - Troubleshooting guide
- Git Diff Troubleshooting - Debug git diff issues
- Server Resilience - Persistence features
- Server Supervisor - Auto-restart configuration
- Get Open Tasks: get_next_tasks
- Monitor progress: work_status taskId="task-123"
- Review changes: review_changes diffId="diff-456"
- Apply changes: apply_changes diffId="diff-456" targetWorkspace="/project"
- Continue: get_next_tasks workFolder="/project"
Server won't start?
parallel-work health # Check prerequisites
parallel-work logs -e # Check error logs
Tasks timing out?
parallel-work follow # Monitor task execution
view_container_logs identifier="task-123"
Git diff issues? See Git Diff Troubleshooting
- Fork the repository
- Create a feature branch
- Make your changes
- Test it yourself!
- Submit a pull request
MIT License - see LICENSE file for details.
