| name | incident-intake |
| description | Normalize incident reports into a structured brief with severity, impact, and immediate context. |
| user-invocable | false |
Incident Intake Skill
Convert raw incident input into a consistent incident brief before triage starts.
Input parameters
| Parameter | Required | Description |
|---|
task_path | Yes | Absolute path to incident task directory |
incident_description | Yes | User-provided incident summary |
Execution workflow
Step 1: Build normalized incident brief
Extract and structure:
- symptom summary,
- affected surface (service, endpoint, region, tenant),
- first-seen timestamp (if known),
- business impact,
- known recent changes/deploys.
If required fields are missing, use ask_user to fill only critical gaps.
Step 2: Assess initial severity
Classify tentative severity:
SEV-1: major outage / critical business impact,
SEV-2: significant degradation,
SEV-3: contained issue with limited blast radius.
Step 3: Write artifacts
Create:
{task_path}/analysis/incident-brief.md
{task_path}/analysis/impact-assessment.md
Step 4: Return structured output
status: success|partial|failed
severity: SEV-1|SEV-2|SEV-3
impact_summary: "..."
brief_path: analysis/incident-brief.md
impact_path: analysis/impact-assessment.md
Output requirements
incident-brief.md must include:
- Incident title
- Symptoms
- Scope and impact
- Known context and assumptions
- Unknowns to resolve in triage