name: system-monitor
version: 1.0.0
description: System monitoring and performance tuning — CPU, memory, disk, network, and process metrics. Triggered when user asks about system health or performance.
layer: system
lifecycle: operations
category: monitor-perf
tags: [monitoring, performance, cpu, memory, disk, network, metrics]
platforms: [khy-quant, claude-code, cosh]
dependencies: []

System Monitor Skill#

Real-time system health monitoring and performance diagnostics.

When to Activate#

User asks about system resource usage (CPU, memory, disk)
User reports slow performance or high load
User wants to monitor a specific process
User asks about network throughput or connections
User wants to set up alerting thresholds

Monitoring Pipeline#

Step 1: Quick Health Check#

# One-liner system overview
uptime && free -h && df -h / && echo "---" && ps aux --sort=-%mem | head -6

Step 2: Deep Diagnostics#

CPU Analysis#

# Per-core utilization
mpstat -P ALL 1 3 2>/dev/null || top -bn1 | head -5
# Top CPU consumers
ps aux --sort=-%cpu | head -10

Memory Analysis#

# Detailed memory breakdown
free -h
cat /proc/meminfo | head -20 2>/dev/null
# Top memory consumers
ps aux --sort=-%mem | head -10

Disk I/O#

# I/O statistics
iostat -x 1 3 2>/dev/null || df -h
# Find large files
du -sh /home/*/Khy-OS* 2>/dev/null

Network#

# Active connections
ss -tunp | head -20
# Network interface stats
ip -s link 2>/dev/null || netstat -i

Step 3: KHY-Quant Specific#

# Node.js process health
pgrep -af "node.*khy" | head -5
# Backend API response time
curl -o /dev/null -s -w "%{time_total}s" http://localhost:3000/api/health 2>/dev/null
# Database connection check
ls -la ~/.khyquant/data/*.db 2>/dev/null

Step 4: Recommendations#

Based on findings, suggest:

Resource bottlenecks and fixes
Process optimization (restart, reconfigure)
Capacity planning advice

Alert Thresholds#

CPU > 80% sustained: warning
Memory > 85%: warning
Disk > 90%: critical
Load average > 2x CPU cores: warning

Cross-Platform Notes#

Linux: full /proc, systemd, ss support
macOS: use vm_stat, diskutil, lsof
Windows: use Get-Process, Get-Counter, wmic

Safety Rules#

Read-only operations only — never kill processes without confirmation
Respect user privacy — do not inspect process arguments that may contain secrets
Use -n / --no-dns flags to avoid slow DNS lookups