linear-escalation-triagelisted

# Linear Escalation Triage ## Overview Workflow for Linear tickets assigned to you, in two modes: - **Oncall mode** — weekly rotation; proactively handle the queue - **Project mode** — outside oncall; work a specific Linear-tracked ticket Both modes share the same per-ticket protocol. They differ in *entry conditions* and *exit conditions*. **Core principle:** sweep before you dive. Without a sweep, whatever's loudest wins — which is how P1s sit in `Triage` for weeks before someone notices. ## Required tools - `mcp__linear-server__list_issues` — queue sweep - `mcp__linear-server__get_issue` / `list_comments` — pull context - `mcp__linear-server__list_issue_statuses` — enumerate your team's status names (don't assume) - `mcp__linear-server__save_comment` / `save_issue` — post reply, transition status - `Bash(gh:*)` — pull linked external issues, search PR history ## Optional tools (graceful degradation) - `mcp__rsry__*` — bead-based record-keeping (recommended; if unavailable, keep notes locally and skip the `Record` step) - `WebFetch` — for fetching external runbooks / docs - Knowledge-base search (Glean / equivalent) — for prior-art lookups ## When to Use - Start of weekly oncall rotation → full sweep - Specific escalation comes in (Slack ping, external ping, ticket assignment) → per-ticket protocol - Aged P1/P2 in `Triage` > 7 days - Stale-Todo cleanup pass ## When NOT to Use - One-off investigation with no Linear ticket → generic investigation - Code work wher