agent-v3-integration-architect
Agent skill for v3-integration-architect - invoke with $agent-v3-integration-architect
Install with Tessl CLI
npx tessl i github:ruvnet/claude-flow --skill agent-v3-integration-architect31
Does it follow best practices?
If you maintain this skill, you can automatically optimize it using the tessl CLI to improve its score:
npx tessl skill review --optimize ./path/to/skillValidation for skill structure
Evaluation results
100%
96%Agent Framework Modernization: Core Adapter Layer
ClaudeFlowAgent adapter & SONA/Flash Attention integration
agentic-flow import
0%
100%
Extends AgenticFlowAgent
0%
100%
executeWithSONA task handler
0%
100%
legacyCompatibilityLayer method
50%
100%
All 5 SONA modes
0%
100%
SONA setMode call
0%
100%
configureAdaptationRate call
0%
100%
Flash Attention API call
0%
100%
Flash Attention speedup target
0%
100%
Flash Attention memory reduction
0%
100%
Flash Attention mechanisms
0%
100%
Without context: $0.3862 · 1m 52s · 16 turns · 21 in / 6,583 out tokens
With context: $0.5050 · 2m 28s · 19 turns · 539 in / 8,410 out tokens
95%
41%Orchestration Framework Consolidation
Three-phase migration plan & backward-compatible code reduction
Three distinct phases
100%
100%
AgenticFlowAdapter class
50%
100%
migrateSwarmCoordination method
0%
37%
migrateAgentManagement method
0%
100%
Task graph migration
0%
100%
Session migration
50%
100%
Removes SwarmCoordinator.ts
100%
100%
Removes AgentManager.ts
100%
100%
Removes TaskScheduler.ts
50%
100%
Code reduction target
100%
100%
Dual operation stage
100%
100%
validateFullParity before deprecation
0%
100%
Without context: $1.1306 · 5m 52s · 25 turns · 32 in / 25,005 out tokens
With context: $1.1716 · 5m 26s · 29 turns · 446 in / 21,587 out tokens
84%
78%Shared Intelligence Layer for Multi-Agent Platform
AgentDB cross-agent memory & MCP tools/hooks integration
HNSW index type
30%
100%
1536 dimensions
0%
100%
Speedup target reference
0%
100%
enableCrossAgentSharing call
0%
100%
MCP getAvailableTools call
0%
100%
213 tools reference
0%
100%
registerClaudeFlowSpecificTools call
0%
0%
hooks.getTypes call
0%
100%
19 hook types reference
0%
100%
configureClaudeFlowHooks call
0%
0%
RL training via agenticFlow.rl.train
0%
100%
At least 3 named RL algorithms
50%
100%
Without context: $0.8507 · 4m 24s · 20 turns · 27 in / 19,894 out tokens
With context: $0.7959 · 3m 31s · 27 turns · 31 in / 13,330 out tokens
- Evaluated
- Agent
- Claude Code
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.