Escalation

---
name: escalation
description: "Guide support teams on when and how to escalate issues effectively with structured briefs"
type: reference
version: 1.0.0
category: business
last_updated: 2026-02-03
source: https://github.com/anthropics/knowledge-work-plugins
related_skills:
  - ticket-triage
  - customer-research
  - response-drafting
  - knowledge-management
capabilities: []
requires: []
see_also: []
tags: []
---

# Escalation Skill Summary

This guide teaches support teams when and how to escalate issues effectively. Here are the key takeaways:

## Escalation Triggers

Escalate when issues involve confirmed bugs, production outages, data integrity risks, security concerns, or high-value customers at risk. Technical escalations include: bug confirmed and needs a code fix, infrastructure investigation needed, data corruption or loss.

## Tier Structure

Issues flow through escalation levels (L1 -> L2 -> Engineering/Product) based on complexity and impact. Security concerns bypass normal progression and go immediately to the security team.

## Critical Success Factor: Reproduction Steps

Good reproduction steps are the single most valuable thing in a bug escalation. Effective steps start from a clean state, use specific values, document the environment, and include exact error messages.

## Quantifying Impact

Strong escalations measure impact across breadth (how many customers), depth (severity), duration, revenue at risk, reputation risk, and contractual obligations. This prevents vague escalations from being deprioritized.

## Post-Escalation Ownership

Support maintains customer relationship responsibility after escalating. Follow-up frequency depends on severity: critical issues get updates every 2-4 hours; high-severity issues every 4-8 hours; medium issues every 1-2 business days.