---

name: system-health description: Show agent-mobile system health when user asks "health", "system health", "check health", or runs /health allowed-tools:

• Bash

---

System Health Skill

Shows the health of all agent-mobile systems at a glance.

Trigger Phrases

• "health" or "/health"
• "system health"
• "check health"
• "agent health"

Usage

When triggered, run the status script:

Copy$HOME/.claude/skills/system-health/scripts/status.sh

What It Checks

1. Credentials - OAuth token backup exists and is recent
2. Skills Sync - Git status of skills folder (clean/dirty, last commit)
3. Tailscale - Connection status and hostname
4. Notifications - ntfy config and topic
5. Ralph Loops - Active loops with iteration counts
6. Deployed Services - Running Docker containers with uptime and ports
7. Backups - Age of credential, config, and settings backups

IMPORTANT: Check Before Deploying

Always run /health before deploying any services to avoid conflicts with other Claude sessions.

If you see services already running in "Deployed Services", do NOT deploy again unless:

• User explicitly asks to restart/redeploy
• You need to update to a new version

Deploying when services are already running causes:

• Duplicate containers
• Port conflicts
• Database corruption
• Need for manual cleanup

Output Format

Copy╭─────────────────────────────────────────────────────────╮
│  AGENT-MOBILE STATUS                                    │
╰─────────────────────────────────────────────────────────╯

Credentials:     ✓ Valid (backup: 5 min ago)
Skills Sync:     ✓ Clean (last commit: 2h ago)
Tailscale:       ✓ Connected (agent-mobile.tail1234.ts.net)
Notifications:   ✓ Enabled (topic: my-claude-xyz)

Active Ralph Loops:
  • trading (iteration 12/50)
  • zones (iteration 3/20)

Deployed Services:
  ● app-api (up 2h, port 3001)
  ● app-web (up 2h, port 3000)
  ● redis (up 2h, port 6379)

Backups:
  • credentials: 5 min ago
  • config: 5 min ago
  • settings.local: 5 min ago