- 8 New Specialist Agents: Complete development lifecycle coverage
project-planner: Strategic planning and task decompositionapi-developer: Backend API development specialistfrontend-developer: Modern web interface specialisttdd-specialist: Test-driven development expertapi-documenter: Technical documentation specialistdevops-engineer: Infrastructure and deployment expertproduct-manager: Product strategy and planningmarketing-writer: Technical marketing content expert
- 2-4x Performance Improvement: All agents now execute operations concurrently
- Batch Operations: Multiple file reads/writes in single operations
- Parallel Agent Execution: Run multiple agents simultaneously
- Performance Benchmarks: ~80% improvement in multi-agent workflows
- Shared Memory Store: Agents coordinate through persistent memory
- JSON-based Persistence: Lightweight, file-based storage
- TTL Support: Automatic expiration of temporary data
- Namespace Isolation: Organized data storage patterns
- Pattern Matching: Query memory with wildcards
- Real-time Monitoring: Track agent status and performance
- Task Management: View and manage agent tasks
- Memory Viewer: Inspect shared memory contents
- Agent Control: Run agents directly from dashboard
- ShadCN UI: Modern, responsive interface on port 7842
- CLI Mode: Run agents outside Claude Code
claude-agents run marketing-writer --task "Write launch post" - File Input: Process tasks from files
- Interactive Mode: Step-by-step agent interaction
- Automation Ready: Perfect for CI/CD integration
- Event-Driven Automation: Trigger actions on agent events
- Memory Coordination Hooks: React to memory changes
- Task Completion Hooks: Post-processing workflows
- Conditional Execution: Smart hook triggering
- ✅ CLAUDE.md configuration for concurrent execution rules
- ✅ SimpleMemoryStore with JSON persistence
- ✅ Comprehensive test suite (memory, concurrent, commands)
- ✅ Updated all agents with concurrent execution patterns
- ✅ Hooks for all agents with automation workflows
- ✅
claude-agents run <agent>: Independent agent execution - ✅
claude-agents dashboard: Launch web interface - ✅ New slash commands for all specialist agents
- ✅ Hooks system documentation
- ✅ Example workflows guide
- ✅ Agent creation guide
- ✅ Updated README with all new features
- File Operations: 80% faster with concurrent reads
- Multi-Agent Tasks: 80.1% improvement in parallel execution
- Mixed Operations: 34ms average per operation (vs 200ms sequential)
- Memory Coordination: <5ms for inter-agent communication
- Memory Directory: Changed from
.claude-agentsto.swarm - Hook Format: New JSON structure for hook definitions
- Agent Metadata: Enhanced metadata.json format
npm install -g @webdevtoday/claude-agents@latest
claude-agents install --allnpm update -g @webdevtoday/claude-agents
# Reinstall agents to get new features
claude-agents install --all --force- Fixed memory persistence issues
- Resolved concurrent execution race conditions
- Improved error handling in agent execution
- Fixed dashboard API endpoint errors
- Agent marketplace for community sharing
- Cloud memory synchronization
- Advanced analytics and metrics
- Plugin system for agent extensions
- Mobile dashboard companion app
Thanks to all contributors and the Claude Code community for feedback and suggestions!
├── 15 Specialized Agents (7 original + 8 new)
├── Memory System with Persistence
├── Web Dashboard with ShadCN UI
├── Comprehensive Test Suite
├── Example Workflows Documentation
├── Agent Creation Guide
└── Full Concurrent Execution Support
Claude Sub-Agents 2.0 transforms Claude Code into a complete development shop with specialized AI agents covering every aspect of the software development lifecycle, all working together at maximum efficiency through concurrent execution and shared memory.
- Fixed package.json configuration
- SEO optimization
- Bug fixes
- Added agent removal functionality
- Improved status indicators
- Initial release fixes
- Initial release
- 6 core agents
- Basic CLI functionality