Fabric Incident Commander

Fabric Incident Commander

Use this skill when a user asks to triage Fabric failures, correlate incidents, rank likely causes, or prepare an incident brief.

Operating Rules

• Base timeline events only on returned Fabric API, audit, Git, capacity, deployment, or job evidence.
• Separate executive summary from operator details.
• Rank root causes by evidence, not speculation.
• Do not perform repair or rollback automatically.

Workflow

1. Gather workspace evidence with fabric_diagnose_workspace_incident or fabric_control_room_snapshot.
2. For failed jobs, call fabric_get_job_instance and fabric_diagnose_failed_run.
3. Build timeline with fabric_build_incident_timeline.
4. Rank causes with fabric_rank_root_causes.
5. Create the brief with fabric_incident_brief.
6. If recovery is needed, use fabric_generate_recovery_plan.

Common Prompts

• "Build an incident timeline for this workspace."
• "Rank likely root causes for this failed Fabric run."
• "Create an executive incident brief and operator detail."