# Skill: Triage a support request

Use for existing clients reporting something broken, slow or wrong in their HighLevel build.

## Goal

Capture enough detail for a engineer to reproduce the issue — severity, scope, and evidence — without attempting fixes that need production access.

## Steps

1. **Verify client status** — existing client vs new enquiry. New enquiries go to qualify-enquiry, not support.
2. **Severity** — ask explicitly:
   - **Critical** — live customer-facing failure (calls not routing, payments failing, mass send broken).
   - **High** — major workflow broken but workaround exists.
   - **Normal** — incorrect behaviour, needs fix but not burning.
   - **Low** — question, cosmetic, or enhancement.
3. **Scope** — one sub-account or many? One workflow or whole account? Started when?
4. **Evidence** — request: screenshot, Loom link, workflow name, contact ID (not PII beyond what's needed), error message verbatim.
5. **Create ticket** — support pipeline with all fields populated; attach summary note.
6. **Set expectation** — critical = same business day acknowledgment; normal = within SLA defined in knowledge base.

## Guardrails

- Do not access or modify live workflows without human approval.
- Do not ask for passwords, API keys or full account exports in chat.
- Do not diagnose root cause unless documented in KB — describe symptoms only.
- Escalate critical issues to human immediately after ticket creation.

## Ticket summary template

```
Severity: [critical/high/normal/low]
Account: [sub-account name or ID]
Symptom: [what they see]
Expected: [what should happen]
Since: [when it started]
Workflow/tool: [name if known]
Evidence: [links]
Already tried: [their steps]
```