CLI Tools - ruvnet/ruv-FANN GitHub Wiki
The ruv-FANN ecosystem provides powerful command-line interfaces for neural network management, swarm orchestration, and distributed AI operations. This guide covers all available CLI tools and their comprehensive usage.
The primary CLI for distributed swarm orchestration and coordination.
# Install globally
npm install -g ruv-swarm
# Or use with npx
npx ruv-swarm --help# Basic initialization
ruv-swarm init --topology mesh --max-agents 10
# Advanced initialization with custom config
ruv-swarm init \
--topology hierarchical \
--max-agents 20 \
--strategy balanced \
--persistence sqlite \
--neural-acceleration true# Spawn single agent
ruv-swarm spawn researcher --name "research-agent-1"
# Spawn multiple agents in parallel
ruv-swarm spawn \
researcher:3 \
analyst:2 \
coder:2 \
--parallel true# Simple task
ruv-swarm orchestrate "Analyze sales data" --priority high
# Complex task with parameters
ruv-swarm orchestrate \
--task "Build REST API with authentication" \
--priority high \
--params '{"framework": "express", "database": "postgresql"}' \
--timeout 3600# Get swarm status
ruv-swarm status
# Detailed agent status
ruv-swarm status --agents --detailed
# Real-time monitoring
ruv-swarm monitor --interval 5 --output table# View current configuration
ruv-swarm config list
# Set configuration values
ruv-swarm config set max-agents 15
ruv-swarm config set neural.simd true
ruv-swarm config set persistence.type redis
# Load configuration from file
ruv-swarm config load ./swarm-config.json# Initialize SQLite persistence
ruv-swarm persistence init sqlite --path ./swarm.db
# Initialize Redis persistence
ruv-swarm persistence init redis --url redis://localhost:6379
# Backup swarm state
ruv-swarm backup create --output ./backup-$(date +%Y%m%d).json
# Restore from backup
ruv-swarm backup restore ./backup-20250101.jsonUniversal CLI wrapper with hooks system for AI development workflows.
npm install -g claude-flow# Pre-task hook
claude-flow hooks pre-task --description "Start neural training"
# Post-edit hook (after file modifications)
claude-flow hooks post-edit --file "model.py" --memory-key "training/progress"
# Notification hook (for sharing updates)
claude-flow hooks notification --message "Training epoch 50 complete"
# Session management
claude-flow hooks session-end --export-metrics true# Store coordination data
claude-flow memory store "swarm/agent-1/status" "active"
# Retrieve coordination data
claude-flow memory get "swarm/agent-1/status"
# List all memory keys
claude-flow memory list --pattern "swarm/*"# Train neural patterns
claude-flow neural train \
--pattern-type coordination \
--training-data ./coordination-data.json \
--epochs 100
# Get neural status
claude-flow neural status --model-id coordination-model-v1CLI for managing decentralized autonomous agents with blockchain integration.
cargo install daa-cli# Start DAA services
daa service start --config ./daa-config.toml
# Check service status
daa service status --detailed
# Stop services gracefully
daa service stop --timeout 30# Deploy new agent
daa agent deploy \
--name "trading-agent" \
--type economic \
--rules ./trading-rules.json
# List active agents
daa agent list --status active
# Agent performance metrics
daa agent metrics --agent-id agent-123 --timeframe 24h# Add new rule
daa rules add \
--name "risk-management" \
--condition "portfolio_risk > 0.1" \
--action "reduce_position"
# Test rule
daa rules test --rule-id risk-management --dry-run# Join QuDAG network
daa network join --bootstrap-node 192.168.1.100:8001
# Check network status
daa network status --show-peers
# Broadcast transaction
daa network broadcast --tx-data ./transaction.json# Basic WASM build
./build-wasm.sh
# Advanced WASM with SIMD
./build-wasm.sh --features simd --optimize --target web
# Development WASM build
./build-simple-wasm.sh --dev# Run comprehensive benchmarks
./run_benchmarks.sh
# Specific benchmark categories
./run_benchmarks.sh --category simd_performance
./run_benchmarks.sh --category attention_heads
./run_benchmarks.sh --category routing_latency
# Custom benchmark configuration
./run_benchmarks.sh \
--config custom \
--features "std,parallel,simd" \
--output ./results/# Run all tests
cargo test --all-features
# Run specific test suite
cargo test --package micro_core --features simd
# Run benchmarks
cargo bench --features bench
# Generate documentation
cargo doc --all-features --open{
"swarm": {
"maxAgents": 10,
"topology": "hierarchical",
"strategy": "balanced",
"coordinationTimeout": 30000
},
"neural": {
"simdEnabled": true,
"precision": "f32",
"batchSize": 32
},
"persistence": {
"type": "sqlite",
"path": "./swarm.db",
"connectionPoolSize": 10
},
"logging": {
"level": "info",
"format": "json",
"output": "./logs/swarm.log"
}
}{
"hooks": {
"preTask": {
"enabled": true,
"autoSpawnAgents": false,
"memoryPersistence": true
},
"postEdit": {
"enabled": true,
"autoAnalyze": true,
"memoryStorage": true
}
},
"memory": {
"backend": "sqlite",
"path": "./claude-flow-memory.db",
"compression": true
},
"neural": {
"enabled": true,
"modelPath": "./models/",
"autoTrain": true
}
}# Swarm settings
export RUV_SWARM_MAX_AGENTS=10
export RUV_SWARM_USE_SIMD=true
export RUV_SWARM_MEMORY_POOL=2GB
export RUV_SWARM_TOPOLOGY=hierarchical
# Neural settings
export RUV_NEURAL_PRECISION=f32
export RUV_NEURAL_BATCH_SIZE=32
export RUV_NEURAL_USE_GPU=true
# Logging
export RUST_LOG=info
export RUV_LOG_LEVEL=debug
export RUV_LOG_FORMAT=json# Server settings
export RUV_SWARM_HOST=0.0.0.0
export RUV_SWARM_PORT=3001
export RUV_SWARM_WEBSOCKET_PORT=3002
# Database
export RUV_SWARM_DB_URL=sqlite://swarm.db
export RUV_SWARM_DB_POOL_SIZE=10
# Claude Flow
export CLAUDE_FLOW_MEMORY_PATH=./memory.db
export CLAUDE_FLOW_HOOKS_ENABLED=true# Check system resources
ruv-swarm diagnostics --check-resources
# Validate configuration
ruv-swarm config validate
# Reset to defaults
ruv-swarm config reset --confirm# Test network connectivity
ruv-swarm network test --target-agent agent-123
# Check agent status
ruv-swarm agents status --agent-id agent-123 --verbose
# Restart failed agents
ruv-swarm agents restart --failed-only# Performance profiling
ruv-swarm profile --duration 60 --output profile.json
# Memory usage analysis
ruv-swarm memory analyze --show-leaks
# SIMD capability check
ruv-swarm capabilities --simd# Verbose logging
ruv-swarm --log-level debug <command>
# Trace execution
RUST_LOG=trace ruv-swarm <command>
# Memory debugging
RUST_BACKTRACE=1 ruv-swarm <command>
# Performance debugging
ruv-swarm <command> --profile --memory-stats#!/bin/bash
# Automated swarm deployment script
echo "🚀 Deploying production swarm..."
# Initialize swarm with production config
ruv-swarm init \
--config ./production.json \
--topology hierarchical \
--max-agents 50
# Deploy specialized agents
ruv-swarm spawn researcher:10 --batch
ruv-swarm spawn analyst:10 --batch
ruv-swarm spawn coder:20 --batch
ruv-swarm spawn coordinator:5 --batch
ruv-swarm spawn monitor:5 --batch
# Verify deployment
ruv-swarm status --agents --health-check
echo "✅ Swarm deployment complete!"# CI/CD Integration
- name: Deploy AI Swarm
run: |
ruv-swarm init --config ./ci-config.json
ruv-swarm spawn researcher:5 coder:5
ruv-swarm orchestrate "Run test suite" --wait
ruv-swarm status --health-check || exit 1
# Docker Integration
docker run -d \
--name ruv-swarm \
-p 3001:3001 \
-e RUV_SWARM_MAX_AGENTS=20 \
-v ./config:/config \
ruv-swarm:latest \
ruv-swarm start --config /config/swarm.jsonThis comprehensive CLI documentation provides everything needed to effectively use all tools in the ruv-FANN ecosystem for production AI development and deployment.