| name | triage |
| description | Triage production incidents involving data corruption, data loss, slow performance, or outages. Classify severity and recommend immediate actions. |
Incident Report Triage
When a user provides an incident report, follow these steps in order.
Step 1: Classify Severity
Read the incident text and classify:
- Contains "data loss" or "corruption" → Critical → go to Step 2
- Contains "slow" or "timeout" → Degraded → go to Step 3
- Contains "cosmetic" or "typo" → Low → go to Step 4
State the severity classification and transition reason.
Step 2: Critical Incident Response
- State: "This is a Critical incident requiring immediate action."
- Identify the affected component from the report text.
- Recommend: rollback to last known good state.
- Recommend: notify the on-call team.
- Proceed to Step 5.
Step 3: Degraded Incident Response
- State: "This is a Degraded incident requiring investigation."
- Identify the affected component from the report text.
- Recommend: capture diagnostics.
- Recommend: scale up if load-related.
- Proceed to Step 5.
Step 4: Low Incident Response
- State: "This is a Low priority incident."
- Recommend: file a backlog ticket.
- Proceed to Step 5.
Step 5: Summary
Provide a structured summary:
- Severity: (Critical/Degraded/Low)
- Component: (identified from text)
- Actions: (list from the corresponding step)
