| name | fleet-health |
| description | One-page fleet overview — system status, Helm routing, satisfaction scores. |
| version | 1.0.0 |
| author | captain |
| tags | ["fleet","health","overview","status"] |
| intent | Reliable [I05] |
fleet-health
Produce a one-page fleet state summary.
Trigger
/fleet-health
Process
- Call
system_status — gateway health, disk usage, memory, Ollama status.
- Call
provider_health — which API providers are reachable, latency, cooldowns.
- Call
helm_report — routing state, active cooldowns, recent errors.
- Call
satisfaction_scores — fleet average, bottom 3 agents, any alerts.
- Combine into a single-page summary.
MCP Tools Used
system_status — Infrastructure health
provider_health — API provider reachability + cooldowns (zero tokens, ~1s)
helm_report — Engine routing state
satisfaction_scores — Agent satisfaction data
Output Format
# Fleet Health — [date]
## Infrastructure
- Gateway: [status]
- Disk: [usage]
- Memory: [usage]
- Ollama: [status]
## Providers
- Anthropic: [reachable/down] [latency] [cooldown or "ok"]
- Google: [reachable/down] [latency] [cooldown or "ok"]
- OpenAI: [reachable/down] [latency] [cooldown or "ok"]
## Routing
- Engines: [count active]
- Cooldowns: [list or "none"]
- Recent errors: [count or "none"]
## Satisfaction
- Fleet avg: XX/100
- Bottom 3: [agent: score, ...]
- Alerts: [list or "none"]
## Action needed
- [prioritized list or "Fleet healthy"]