name: "V3"
description: "Claude-Flow V3 implementation hub. Routes to the appropriate v3 specialist skill for architecture, security, memory, performance, integration, and swarm coordination."
version: "1.0.0"
category: "development"
tags: ["v3", "claude-flow", "architecture", "hub"]
Claude-Flow V3 Hub
Central routing skill for all V3 implementation work. Use this when you want V3 help without remembering the specific sub-skill name.
V3 Sub-Skills
| Skill |
Domain |
What It Does |
v3-swarm-coordination |
Orchestration |
15-agent hierarchical mesh, phase management |
v3-security-overhaul |
Security |
CVE remediation, threat modeling, secure-by-default |
v3-core-implementation |
Core |
DDD domains, clean architecture, dependency injection |
v3-memory-unification |
Memory |
AgentDB, HNSW indexing, 150x-12,500x search improvement |
v3-integration-deep |
Integration |
agentic-flow@alpha integration, eliminate 10K+ duplicate lines |
v3-performance-optimization |
Performance |
2.49x-7.47x Flash Attention, benchmarking suite |
v3-ddd-architecture |
Architecture |
Bounded contexts, microkernel pattern, clean separation |
v3-mcp-optimization |
MCP |
Connection pooling, load balancing, sub-100ms responses |
v3-cli-modernization |
CLI |
Interactive prompts, hooks integration, workflow automation |
Quick Start
Full Swarm (all 15 agents)
# Initialize complete V3 implementation swarm
Task("V3 swarm init", "Initialize 15-agent hierarchical mesh for v3", "v3-queen-coordinator")
By Domain
# Security first (Phase 1)
Task("V3 security", "CVE remediation and threat modeling", "v3-security-architect")
# Core systems (Phase 2)
Task("V3 core", "DDD architecture implementation", "v3-core-implementation")
Task("V3 memory", "AgentDB unification with HNSW", "v3-memory-specialist")
# Integration (Phase 3)
Task("V3 integration", "Deep agentic-flow@alpha integration", "v3-integration-architect")
# Performance (Phase 4)
Task("V3 performance", "Benchmark validation and optimization", "v3-performance-engineer")
14-Week Timeline
| Phase |
Weeks |
Agents |
Focus |
| Foundation |
1-2 |
#1-6 |
Security + Architecture |
| Core Systems |
3-6 |
#1, #5-9, #13 |
Memory + Swarm + MCP |
| Integration |
7-10 |
#1, #10-14 |
Integration + CLI + Performance |
| Release |
11-14 |
All 15 |
Optimization + Testing + Release |
10 ADRs
- ADR-001: Deep agentic-flow@alpha integration (eliminate duplicates)
- ADR-002: Security-first architecture
- ADR-003: DDD modular architecture
- ADR-004: Unified swarm coordination engine
- ADR-005: MCP server optimization
- ADR-006: Unified memory service (AgentDB)
- ADR-007: CLI modernization with hooks
- ADR-008: SONA learning integration
- ADR-009: Hybrid memory backend (HNSW)
- ADR-010: Performance benchmarking framework
Success Targets
- Performance: 2.49x-7.47x Flash Attention speedup
- Search: 150x-12,500x AgentDB improvement
- Code Size: <5,000 lines (from 15,000+)
- Security Score: 90/100
- Test Coverage: >80%
- Memory: 50-75% reduction
By