mcpField
T
triage-incident

triage-incident

github.com/revitalb-portCloud & DevOps059 installsUpdated 10w ago
Repository

About triage-incident

This skill uses Port MCP context to assess production incident severity and business impact, then draft communications. It is intended for response teams triaging alerts and preparing incidents for investigation.

Key features

  1. 01Gather incident, service, and monitoring context
  2. 02Assess severity from impact criteria
  3. 03Estimate customer and revenue impact
  4. 04Find on-call staff and response teams
  5. 05Draft internal and status-page messages
  6. 06Submit triage recommendations through Port Action

Use cases

  1. 01Triage incidents in the alert stage
  2. 02Prepare incidents for investigation
  3. 03Draft internal and customer-facing updates

MCP support

It queries incident, service, deployment, and team data through Port MCP, then submits triage fields using a Port Action.

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/revitalb-port/skills-registry-demo --skill triage-incident

After installing, type/triage-incidentin Claude Code to confirm it loaded.

Details

From the project README

Triage Incident Skill

Triage production incidents using AI analysis to suggest severity, calculate business impact, and draft communications.

When to Use

  • Incident is in Alert stage and needs triage
  • Analyzing incident severity or impact
  • Preparing incident for investigation phase

Triage Process

Step 1: Gather Context

Collect via Port MCP:

  1. Incident details: Title, description, alert source
  2. Primary service: Tier, dependencies, ownership
  3. Monitoring data: Error rates, affected customers
  4. Timing: Business hours vs off-hours
Step 2: Determine Severity
LevelCriteriaExamples
Sev1Complete tier1 outage, >$10K/hourDatabase crash, auth broken
Sev2Partial tier1 or tier2 outagePayments failing, search down
Sev3Minor degradation, internalSlow dashboard, export broken
Sev4Minimal impact, cosmeticUI typo, styling bug
Step 3: Calculate Business Impact

Provide:

  • Customer impact (estimated users)
  • Revenue impact ($/hour)
  • SLA implications
  • Reputation risk
Step 4: Recommend Response Team

Query Port for:

  1. Service's on_call relation
  2. Service owner relation
  3. Team escalation for sev1/sev2
Step 5: Draft Communications

Internal message:

🚨 Incident: [Title]
Severity: [Level]
Impact: [Brief description]
Response: [Owner] leading, [Team] engaged
ETA: [Timeline]

Status page (customer-facing):

We're experiencing issues with [feature]. 
Our team is actively working on a fix.
We'll provide an update within [time].
Step 6: Submit via Port Action

Call update_incident_triage with:

  • ai_suggested_severity
  • business_impact
  • internal_comms_message
  • status_page_message
  • ai_suggested_owner
  • ai_suggested_response_team

Port MCP Queries

list_entities(blueprint="incident", identifiers=["incident-id"])
list_entities(blueprint="service", include=["on_call", "owner", "team"])
list_entities(blueprint="deployment", query={service: "x", deployed_at: ">1h ago"})

Quality Checklist

  • Severity rationale is data-driven
  • Impact includes metrics
  • Found actual on-call via Port
  • Messages are clear and actionable

Comments

Markdown supported