138+ automated checks across 13 categories. Mighty Mark runs this battery every morning and on demand—catching problems before they become outages.
Every day, Mighty Mark runs the full health check battery against the production fleet. The results feed into Jessie's morning briefing to Greg. A GREEN result means the fleet is healthy. YELLOW flags non-critical issues. RED halts non-essential operations until resolved.
GREEN
All checks pass
YELLOW
Non-critical warnings
RED
Critical — action needed
| Category | Checks | What It Validates |
|---|---|---|
| Gateway | ~10 | Process alive, config valid, plugins loaded, tool count, OpenClaw version |
| Agents | ~8 | Agent configured, system prompts present, knowledge files accessible |
| System | 6 | Disk, memory, CPU, uptime, OOM kills, log file size |
| API | 13 | Anthropic, Telegram, Ethereum RPC, GOG token, Nole platform readiness |
| Data | ~8 | Audit chain integrity, temporal store, backup freshness |
| Memory | 7 | Provider config, embeddings, indexing, staleness, disk usage |
| Fleet | 7 | Fleet cards, audit integrity, bus loaded, message flow, routing |
| Fleet Transport | ~30 | Delivery stats, error rates, ping sweep, guaranteed delivery, circuit breakers |
| Alerting | 11 | Sidecar init, audit trail, behavioral validation, Telegram delivery |
| Security | 9 | Session isolation, exec mode, spawn prevention, secret permissions |
| Temporal | 8 | Noah SQLite, hash chain, cron jobs, fleet corridor, Ethereum anchor |
| Cron | 9 | Schedule completeness, execution freshness, failure rate, dispatch |
| Communication | ~12 | Routing coverage, assessment triple-send, active probes |
# mighty-mark v0.7.x
# Run full health check battery
mighty-mark check
# Run a specific category
mighty-mark check gateway
mighty-mark check security
mighty-mark check fleet
# Run the "house check" — full battery with summary report
mighty-mark check house
# View current fleet status
mighty-mark statusmighty-mark --help for the current surface.Every check function receives a CheckContext that provides access to the OpenClaw gateway, filesystem, API clients, and configuration—without hardcoded dependencies. This makes checks independently testable and allows the runner to inject mock contexts during development.
Each check returns a structured CheckResult with status (PASS/WARN/FAIL), category, message, optional details, and severity classification. Results are aggregated into a report with pass/fail/warn counts per category.
When Mighty Mark detects missing memory directories (embeddings, provider config, indexing state), he creates them automatically. This is the one exception to the read-only constraint—Mighty Mark can create his own operational directories, but never modifies other agents' state.