| name | manage-nextcloud-incidents |
| description | Use when recording or updating durable monitoring state in Nextcloud, including incidents, baselines, and escalation trails. |
Nextcloud Incident Ops
Watchdog uses Nextcloud for durable monitoring state.
Typical Paths
- Nextcloud
/Projects/ai-homebase/incidents/YYYY-MM-DD-short-title.md
- Nextcloud
/Projects/ai-homebase/baselines.md
- Nextcloud
/Projects/ai-homebase/escalation-rules.md
- Nextcloud
/Projects/ai-homebase/watchdog-status-log.md
- Nextcloud
/Projects/ai-homebase/coordination-status.json
Procedure
- Treat Nextcloud paths as remote paths.
- For routine all-clear checks, update only Nextcloud
/Projects/ai-homebase/watchdog-status-log.md when anything needs to be preserved. Do not create a new document.
- For meaningful incidents or baseline changes, write the smallest durable artifact that preserves the evidence.
- Update Nextcloud
/Desk/current.md or Nextcloud /Desk/daily/ when another session should be aware of an active or recent incident quickly.
- If you create a recurring shared monitoring surface outside the usual incident paths, register it in Nextcloud
/Desk/index.md.
- After writing an incident report, store a Qdrant summary with
project, tags, and nc_refs.
- Escalate to main only when the severity gate is met.
- Keep per-check scratch work local; only promote durable patterns, baselines, and incident state.
Use classify-severity-and-escalate to decide whether the case is info, warning, or critical.
Boundaries
- Prefer one evolving baseline or status log over many fragmented notes.
- Do not turn Nextcloud
/Desk/ into an incident archive; it is for bounded continuity.
- Keep incident reports factual and concise.
