/sc-heartbeat

/sc-heartbeat - Quick Heartbeat Report

Run all enabled collectors and display a compact system health report.

Activation

Triggered by: /sc-heartbeat, "heartbeat", "system health", "how is the system", "quick check"

Steps

1. Run system collector: Gather core system metrics via Bash:

   # CPU usage
   top -l 1 -n 0 | grep "CPU usage"
   
   # Memory usage
   memory_pressure | head -1
   # or: vm_stat | head -5
   
   # Disk usage
   df -h / | tail -1
   

   Format into metrics:

   • CPU: usage percentage and load average
   • Memory: used/total and pressure level
   • Disk: used/total and percentage

2. Run dev collector: Check development environment health:

   # Git status of current project
   git -C ~/superclaw status --short | wc -l
   
   # Node.js processes
   pgrep -c node
   
   # Recent build status (check if bridge files are fresh)
   ls -la ~/superclaw/bridge/*.cjs 2>/dev/null | awk '{print $6, $7, $8, $9}'
   

3. Check thresholds: Compare metrics against default thresholds:

   Metric	Warning	Critical
   CPU	> 70%	> 90%
   Memory	> 75%	> 90%
   Disk	> 80%	> 95%

   Flag any metrics that exceed thresholds.

4. Save heartbeat: Write the report to ~/superclaw/data/heartbeats/ with a timestamped filename:

   ~/superclaw/data/heartbeats/heartbeat-2026-02-12T10-30-00.json
   

5. Format compact report: Display results in a concise format:

   Heartbeat Report - 2026-02-12 10:30:00
   ----------------------------------------
   CPU:    23% (load: 2.1, 1.8, 1.5)     OK
   Memory: 12.4/16 GB (78%)              WARN
   Disk:   180/500 GB (36%)              OK
   ----------------------------------------
   Processes: 4 node | Git: 2 changes
   Bridges:   3/3 built (12 min ago)
   ----------------------------------------
   Alerts: 1 warning (memory > 75%)
   

6. Alert if critical: If any metric exceeds the critical threshold, send an alert to Telegram (if configured):

   sc_send_message({
     channel: "telegram",
     message: "CRITICAL: Memory at 92% on MacBook"
   })
   

Error Handling

• If system commands fail (e.g., top not available), skip that metric and note it as "unavailable"
• If heartbeat directory does not exist, create it before saving
• If Telegram is not configured, skip the alert step (still show the report locally)
• If previous heartbeat file cannot be written, still display the report to the user

Notes

• This is a quick one-shot check. For recurring heartbeats, use the cron-mgr skill to schedule this.
• The full heartbeat skill provides more collectors (GitHub, Sentry, calendar). This command runs only system and dev collectors for speed.
• Heartbeat history is stored as JSON files for trend analysis.