triage-a-ticket

Triage a Ticket

When to use

New inbound message landed, no conversations.json entry for thread yet, OR existing entry need re-triage because content changed (e.g. how-to turned outage report). Solo founder: triage constant - every fresh reply need this skill run first.

Connections I need

I run external work through Composio. Before this skill runs I check that the categories below are linked. Missing → I name the category, ask you to connect it from the Integrations tab, stop.

• Inbox (Gmail / Outlook) - fetch the inbound message and full thread. Required.
• Support helpdesk (Intercom / Zendesk / Help Scout) - alternate source if customer messages land there. Required if you don't use Gmail / Outlook for support.
• Messaging (Slack) - source of customer DMs you triage as tickets. Optional.

If no inbox or helpdesk is connected I stop and ask you to link the one your customers actually message.

Information I need

I read your support context first. For every required field that's missing I ask ONE plain-language question (best modality: connected app > file drop > URL > paste) and wait.

• Routing categories - Required. Why I need it: I sort every inbound into one. If missing I ask: "When a ticket comes in, what buckets do you sort it into - bug, how-to, billing, anything else?"
• Response-time tiers - Required. Why I need it: priority assignment depends on tier thresholds. If missing I ask: "What response time do you want to hit for your most urgent tickets, and what's acceptable for the rest?"
• VIP list - Required. Why I need it: VIPs floor at P1 regardless of content. If missing I ask: "Which 3 to 5 customers should always get top priority?"
• Monthly revenue / plan tier per customer - Optional. Why I need it: lets me weight priority by paying-customer status. If you don't have it I keep going with TBD and weight by content signals only.

Steps

1. Read context/support-context.md. If missing, stop. Tell me run set-up-my-support-info first. Read routing rules + response-time tiers + VIP list from doc - never hardcode.
2. Identify source - you name channel or message referenced by external id. Use composio search <channel> to find correct fetch slug (e.g. Gmail thread fetch, Intercom conversation fetch). Do NOT hardcode tool slugs.
3. Fetch raw thread via Composio. Pull subject, all messages, sender email, external message ids.
4. Resolve customer. Look up customers.json by sender email. If not found, create new index entry (slug = kebab-cased email local-part, deduped if needed).
5. Categorize content against routing categories in context/support-context.md (typical set: bug | how-to | feature | billing | account | security | other). Content signals: error messages + stack traces lean bug; "how do I…" lean how-to; "can you add…" lean feature; keywords "refund", "invoice", "charge" lean billing.
6. Assign priority (P1–P4) using tier thresholds from context/support-context.md. Typical start rules: monthly revenue >= $500/mo → base P2; VIP tag → P1 floor. Escalate on content: "down", "can't log in", "data loss", "production" → bump one level (max P1). De-escalate on "whenever you get a chance".
7. Set response-time fields using domains.inbox.responseTimeTargets.firstResponseHours (fallback to tier table in context doc). breached = false initially.
8. Write atomically. Upsert into conversations.json. Write full messages to conversations/{id}/thread.json.
9. Append to outputs.json with type: "triage", domain: "inbox", title = {customer} - {subject}, summary = category + priority, path.